Skip to content

Force FP32 compute in GLM4 FFN Down #13101

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Apr 25, 2025
Merged

Force FP32 compute in GLM4 FFN Down #13101

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6efd872
Force FP32 compute in cuBLAS GEMM
city96 Apr 24, 2025
db52579
Revert "Force FP32 compute in cuBLAS GEMM"
city96 Apr 25, 2025
7097567
Force F32 compute in GLM4 ffn down
city96 Apr 25, 2025
06113f0
Edit comment to clarify issue
city96 Apr 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 11 additions & 29 deletions ggml/src/ggml-cuda/ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1245,37 +1245,19 @@ static void ggml_cuda_op_mul_mat_cublas(
}
const half * src1_ptr = src1->type == GGML_TYPE_F16 ? (const half *) src1_ddf_i : src1_as_f16.get();

CUBLAS_CHECK(cublasSetStream(ctx.cublas_handle(id), stream));

if (GGML_CUDA_CC_IS_CDNA(cc) || GGML_CUDA_CC_IS_RDNA4(cc)) {
const float alpha = 1.0f;
const float beta = 0.0f;
CUBLAS_CHECK(
cublasGemmEx(ctx.cublas_handle(id), CUBLAS_OP_T, CUBLAS_OP_N,
row_diff, src1_ncols, ne10,
&alpha, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&beta, dst_dd_i, CUDA_R_32F, ldc,
CUBLAS_COMPUTE_32F,
CUBLAS_GEMM_DEFAULT_TENSOR_OP));
} else {
ggml_cuda_pool_alloc<half> dst_f16(ctx.pool(id), row_diff*src1_ncols);

const half alpha_f16 = 1.0f;
const half beta_f16 = 0.0f;
const float alpha = 1.0f;
const float beta = 0.0f;

CUBLAS_CHECK(
cublasGemmEx(ctx.cublas_handle(id), CUBLAS_OP_T, CUBLAS_OP_N,
row_diff, src1_ncols, ne10,
&alpha_f16, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&beta_f16, dst_f16.get(), CUDA_R_16F, ldc,
CUBLAS_COMPUTE_16F,
CUBLAS_GEMM_DEFAULT_TENSOR_OP));
CUBLAS_CHECK(cublasSetStream(ctx.cublas_handle(id), stream));
CUBLAS_CHECK(
cublasGemmEx(ctx.cublas_handle(id), CUBLAS_OP_T, CUBLAS_OP_N,
row_diff, src1_ncols, ne10,
&alpha, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&beta, dst_dd_i, CUDA_R_32F, ldc,
CUBLAS_COMPUTE_32F,
CUBLAS_GEMM_DEFAULT_TENSOR_OP));

const to_fp32_cuda_t to_fp32_cuda = ggml_get_to_fp32_cuda(GGML_TYPE_F16);
to_fp32_cuda(dst_f16.get(), dst_dd_i, row_diff*src1_ncols, stream);
}
} else {
ggml_cuda_pool_alloc<float> src0_ddq_as_f32(ctx.pool(id));
ggml_cuda_pool_alloc<float> src1_ddq_as_f32(ctx.pool(id));
Expand Down