Skip to content

perf(ggml/hip): add batched ROCmFPX MMQ #513

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

Draft
weicj wants to merge 1 commit into
base: codex/ds4-rocmfpx-server
Choose a base branch
from
Draft

perf(ggml/hip): add batched ROCmFPX MMQ #513

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
15 changes: 15 additions & 0 deletions server/deps/llama.cpp/ggml/src/ggml-cuda/mmq.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,15 @@ static void ggml_cuda_mul_mat_q_switch_type(ggml_backend_cuda_context & ctx, con
case GGML_TYPE_Q8_0:
mul_mat_q_case<GGML_TYPE_Q8_0>(ctx, args, stream);
break;
case GGML_TYPE_Q4_0_ROCMFP4_FAST:
mul_mat_q_case<GGML_TYPE_Q4_0_ROCMFP4_FAST>(ctx, args, stream);
break;
case GGML_TYPE_Q2_0_ROCMFP2:
mul_mat_q_case<GGML_TYPE_Q2_0_ROCMFP2>(ctx, args, stream);
break;
case GGML_TYPE_Q3_0_ROCMFPX:
mul_mat_q_case<GGML_TYPE_Q3_0_ROCMFPX>(ctx, args, stream);
break;
case GGML_TYPE_MXFP4:
case GGML_TYPE_NVFP4:
#ifndef GGML_CUDA_BLACKWELL_CONSUMER
Expand Down Expand Up @@ -316,6 +325,12 @@ bool ggml_cuda_should_use_mmq(enum ggml_type type, int cc, int64_t ne11, int64_t
case GGML_TYPE_IQ4_NL:
mmq_supported = true;
break;
case GGML_TYPE_Q2_0_ROCMFP2:
case GGML_TYPE_Q3_0_ROCMFPX:
case GGML_TYPE_Q4_0_ROCMFP4_FAST:
// ROCmFPX MMQ variants are implemented for gfx1151 only.
mmq_supported = GGML_CUDA_CC_IS_RDNA3_5(cc);
break;
default:
mmq_supported = false;
break;
Expand Down
Loading