EngineX-Ascend/enginex-ascend-910-llama.cpp
10
0
Fork 0
Code Issues Pull Requests Actions 4 Projects Releases Wiki Activity

metal : Add template specialization for mul_mm_id w/ ne20 == 10 (#15799)

Browse Source 
Branch: GGMLMetalNE20

Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
This commit is contained in:
Gabe Goodhart
2025-09-04 09:53:22 -06:00
committed by GitHub
parent d1e2adba65
commit 856ed0947f
2 changed files with 4 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
ggml/src/ggml-metal/ggml-metal.m
View File

@@ -407,6 +407,7 @@ enum ggml_metal_kernel_type {
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_4,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_6,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_8,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_10,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_16,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F32_F16,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F16_F16,
@@ -1439,6 +1440,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_4, mul_mm_id_map0_f16_ne20_4, has_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_6, mul_mm_id_map0_f16_ne20_6, has_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_8, mul_mm_id_map0_f16_ne20_8, has_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_10, mul_mm_id_map0_f16_ne20_10, has_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_16, mul_mm_id_map0_f16_ne20_16, has_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F32_F16, mul_mm_id_f32_f16, has_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F16_F16, mul_mm_id_f16_f16, has_simdgroup_mm);
@@ -3979,6 +3981,7 @@ static int ggml_metal_encode_node(
case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_4 ].pipeline; break;
case 6: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_6 ].pipeline; break;
case 8: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_8 ].pipeline; break;
case 10: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_10].pipeline; break;
case 16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_16].pipeline; break;
default: GGML_ABORT("missing specialization for ne20 = %d", (int) ne20);
}