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

opencl: add q8_0 mm support (#16469)

Browse Source 
* opencl: add mm_q8_0_f32

* opencl: fix data loading for incomplete tile

* opencl: use q8_0 mm for larger matrix

* opencl: add some tests to cover the path
This commit is contained in:
lhez
2025-10-15 10:51:04 -07:00
committed by GitHub
parent d93f8439b0
commit 0cb7a0683b
6 changed files with 271 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
ggml/src/ggml-opencl/ggml-opencl.cpp
View File

@@ -408,6 +408,7 @@ struct ggml_backend_opencl_context {
cl_program program_mul_mv_id_mxfp4_f32_flat;
cl_program program_mul_mm_f32_f32_l4_lm;
cl_program program_mul_mm_f16_f32_l4_lm;
cl_program program_mul_mm_q8_0_f32_l4_lm;
cl_kernel kernel_add, kernel_add_row, kernel_add_f16, kernel_add_row_f16;
cl_kernel kernel_mul, kernel_mul_row, kernel_mul_f16, kernel_mul_row_f16;
@@ -480,6 +481,7 @@ struct ggml_backend_opencl_context {
cl_kernel kernel_mul_mv_id_mxfp4_f32_flat;
cl_kernel kernel_mul_mm_f32_f32_l4_lm;
cl_kernel kernel_mul_mm_f16_f32_l4_lm;
cl_kernel kernel_mul_mm_q8_0_f32_l4_lm;
std::vector<ProfilingInfo> profiling_info;
@@ -1191,6 +1193,22 @@ static void load_cl_kernels(ggml_backend_opencl_context *backend_ctx, ggml_cl_ve
GGML_LOG_CONT(".");
}
// mul_mm_q8_0_f32_l4_lm
{
#ifdef GGML_OPENCL_EMBED_KERNELS
const std::string kernel_src {
#include "mul_mm_q8_0_f32_l4_lm.cl.h"
};
#else
const std::string kernel_src = read_file("mul_mm_q8_0_f32_l4_lm.cl");
#endif
backend_ctx->program_mul_mm_q8_0_f32_l4_lm =
build_program_from_source(backend_ctx->context, backend_ctx->device, kernel_src.c_str(), compile_opts);
CL_CHECK((backend_ctx->kernel_mul_mm_q8_0_f32_l4_lm = clCreateKernel(backend_ctx->program_mul_mm_q8_0_f32_l4_lm, "kernel_mul_mm_q8_0_f32_l4_lm", &err), err));
GGML_LOG_CONT(".");
}
// mul
{
#ifdef GGML_OPENCL_EMBED_KERNELS
@@ -6961,6 +6979,44 @@ static void ggml_cl_mul_mat(ggml_backend_t backend, const ggml_tensor * src0, co
backend_ctx->enqueue_ndrange_kernel(kernel, 3, global_work_size, local_work_size, dst);
return;
}
case GGML_TYPE_Q8_0: {
if (ne11 < 32) {
break;
}
kernel = backend_ctx->kernel_mul_mm_q8_0_f32_l4_lm;
nth0 = 128; // calculated as (BM*BN)/(TM*TN)
int batch_stride_a = ne00*ne01;
int batch_stride_b = ne10*ne11;
int batch_stride_d = ne0*ne1;
CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0_q8_0->q));
CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &extra0_q8_0->d));
CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device));
CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1));
CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device));
CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd));
CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00));
CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01));
CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02));
CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne11));
CL_CHECK(clSetKernelArg(kernel, 10, sizeof(int), &ne12));
CL_CHECK(clSetKernelArg(kernel, 11, sizeof(int), &ne10)); // stride_a
CL_CHECK(clSetKernelArg(kernel, 12, sizeof(int), &ne10)); // stride_b
CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &ne01)); // stride_d
CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &batch_stride_a));
CL_CHECK(clSetKernelArg(kernel, 15, sizeof(int), &batch_stride_b));
CL_CHECK(clSetKernelArg(kernel, 16, sizeof(int), &batch_stride_d));
CL_CHECK(clSetKernelArg(kernel, 17, sizeof(int), &r2));
CL_CHECK(clSetKernelArg(kernel, 18, sizeof(int), &r3));
// 64 is block tile size BM and BN - change here when BM and BN in the kernel are changed.
size_t global_work_size[] = {(size_t)(CEIL_DIV(ne01, 64)*nth0), (size_t)(CEIL_DIV(ne11, 64)), (size_t)ne12*ne13};
size_t local_work_size[] = {(size_t)nth0, 1, 1};
backend_ctx->enqueue_ndrange_kernel(kernel, 3, global_work_size, local_work_size, dst);
return;
}
default:
break;
}