EngineX-Mthreads/enginex-mthreads-vllm
4
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Sync from v0.13

Browse Source
This commit is contained in:
xiezhongtao
2026-01-19 10:38:50 +08:00
parent b2ef04d792
commit 5aef6c175a
3714 changed files with 854317 additions and 89342 deletions
Download Patch File Download Diff File Expand all files Collapse all files

146
vllm/model_executor/layers/quantization/utils/configs/N=10240,K=5120,device_name=NVIDIA_L40S,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=12288,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=12288,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"1024": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=1536,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_L20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 2
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 2
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"96": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"128": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"256": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2112,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"1024": {
"BLOCK_SIZE_M": 256,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 5
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2112,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H20,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_L20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2048": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"1024": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4096": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=256,K=7168,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 5
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 2
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=NVIDIA_H20,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=1536,device_name=NVIDIA_L20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 2
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H20,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 5
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_L20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 2
},
"4": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"128": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 2
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"256": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 5
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 5
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 2
},
"128": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H20,dtype=int8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 2
},
"48": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"96": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 4
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 256,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_L20,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
},
"4096": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 2
}
}

26
vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_L20Y,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,26 @@
{
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"8": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"24": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=AMD_Instinct_MI325X,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

164
vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=AMD_Instinct_MI325_OAM,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,164 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 16,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"16": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"24": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"48": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"96": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 1,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"1536": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"GROUP_SIZE_M": 8,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"3072": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
},
"4096": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"kpack": 1,
"matrix_instr_nonkdim": 16,
"num_warps": 4
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"2": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"4": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"24": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 2
},
"48": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 2
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 2
},
"512": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 8,
"num_stages": 2
},
"1024": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 8,
"num_stages": 2
},
"1536": {
"BLOCK_SIZE_M": 128,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 8,
"num_stages": 2
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 2
}
}

146
vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128,128].json Normal file
View File

@@ -0,0 +1,146 @@
{
"1": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 5
},
"2": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 5
},
"4": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 4
},
"8": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"16": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"24": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"32": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 4
},
"48": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"64": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 4
},
"96": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"128": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 4
},
"256": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 64,
"num_warps": 4,
"num_stages": 3
},
"1536": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
},
"2048": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"3072": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
},
"4096": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 16,
"num_warps": 4,
"num_stages": 3
}
}

Some files were not shown because too many files have changed in this diff Show More