config: Update fused moe config (#4493)
This commit is contained in:
@@ -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": 3
|
||||||
|
},
|
||||||
|
"2": {
|
||||||
|
"BLOCK_SIZE_M": 64,
|
||||||
|
"BLOCK_SIZE_N": 64,
|
||||||
|
"BLOCK_SIZE_K": 128,
|
||||||
|
"GROUP_SIZE_M": 32,
|
||||||
|
"num_warps": 4,
|
||||||
|
"num_stages": 4
|
||||||
|
},
|
||||||
|
"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": 64,
|
||||||
|
"num_warps": 4,
|
||||||
|
"num_stages": 3
|
||||||
|
},
|
||||||
|
"24": {
|
||||||
|
"BLOCK_SIZE_M": 64,
|
||||||
|
"BLOCK_SIZE_N": 128,
|
||||||
|
"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": 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": 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": 64,
|
||||||
|
"BLOCK_SIZE_N": 128,
|
||||||
|
"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": 64,
|
||||||
|
"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": 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": 3
|
||||||
|
},
|
||||||
|
"1536": {
|
||||||
|
"BLOCK_SIZE_M": 128,
|
||||||
|
"BLOCK_SIZE_N": 64,
|
||||||
|
"BLOCK_SIZE_K": 128,
|
||||||
|
"GROUP_SIZE_M": 1,
|
||||||
|
"num_warps": 4,
|
||||||
|
"num_stages": 4
|
||||||
|
},
|
||||||
|
"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": 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
|
||||||
|
}
|
||||||
|
}
|
||||||
@@ -1,11 +1,11 @@
|
|||||||
{
|
{
|
||||||
"1": {
|
"1": {
|
||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 64,
|
"BLOCK_SIZE_N": 32,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 1,
|
"GROUP_SIZE_M": 16,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 5
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
"2": {
|
"2": {
|
||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
@@ -35,7 +35,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 64,
|
"BLOCK_SIZE_N": 64,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 16,
|
"GROUP_SIZE_M": 32,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -43,7 +43,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 64,
|
"BLOCK_SIZE_N": 64,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 64,
|
"GROUP_SIZE_M": 32,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -51,7 +51,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 128,
|
"BLOCK_SIZE_N": 128,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 32,
|
"GROUP_SIZE_M": 64,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -59,7 +59,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 128,
|
"BLOCK_SIZE_N": 128,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 32,
|
"GROUP_SIZE_M": 64,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -67,7 +67,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 128,
|
"BLOCK_SIZE_N": 128,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 64,
|
"GROUP_SIZE_M": 32,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -91,7 +91,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 128,
|
"BLOCK_SIZE_N": 128,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 32,
|
"GROUP_SIZE_M": 64,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -99,7 +99,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 128,
|
"BLOCK_SIZE_N": 128,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 16,
|
"GROUP_SIZE_M": 32,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
},
|
},
|
||||||
@@ -139,7 +139,7 @@
|
|||||||
"BLOCK_SIZE_M": 64,
|
"BLOCK_SIZE_M": 64,
|
||||||
"BLOCK_SIZE_N": 128,
|
"BLOCK_SIZE_N": 128,
|
||||||
"BLOCK_SIZE_K": 128,
|
"BLOCK_SIZE_K": 128,
|
||||||
"GROUP_SIZE_M": 16,
|
"GROUP_SIZE_M": 32,
|
||||||
"num_warps": 4,
|
"num_warps": 4,
|
||||||
"num_stages": 3
|
"num_stages": 3
|
||||||
}
|
}
|
||||||
|
|||||||
Reference in New Issue
Block a user