mikhail/llama.cpp-mtp-turboquant
1
0
Fork 0
Code Issues Pull Requests Actions 29 Packages Projects Releases Wiki Activity

CUDA: Fix loop unrolling for BW in mul_mat_q_stream_k_fixup (#19053)

Browse Source 
By providing stride_* variables as size_t (i.e., 64-bit) the compiler can
correctly unroll the [two for-loops](https://github.com/ggml-org/llama.cpp/blob/557515be1e93ed8939dd8a7c7d08765fdbe8be31/ggml/src/ggml-cuda/mmq.cuh#L3789-L3816)
on BW. This gives some perf for prefill/pp phase on BW, while not affecting
other SMs:

| GPU                                                     | Model                 | Test   |   t/s master |   t/s osimons/fix_bw_mmq_fixup_kernel |   Speedup |
|:--------------------------------------------------------|:----------------------|:-------|-------------:|--------------------------------------:|----------:|
| NVIDIA RTX 6000 Ada Generation                          | gpt-oss 20B MXFP4 MoE | pp8096 |      8404.05 |                               8375.79 |      1.00 |
| NVIDIA RTX 6000 Ada Generation                          | llama 3B Q4_K_M       | pp8096 |     16148.93 |                              16019.60 |      0.99 |
| NVIDIA RTX 6000 Ada Generation                          | llama 8B Q4_0         | pp8096 |      8008.29 |                               7978.80 |      1.00 |
| NVIDIA RTX 6000 Ada Generation                          | nemotron_h 9B BF16    | pp8096 |      4263.16 |                               4248.53 |      1.00 |
| NVIDIA RTX 6000 Ada Generation                          | nemotron_h 9B Q4_K_M  | pp8096 |      5165.11 |                               5157.43 |      1.00 |
| NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | gpt-oss 20B MXFP4 MoE | pp8096 |     12582.80 |                              12758.37 |      1.01 |
| NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | llama 3B Q4_K_M       | pp8096 |     16879.10 |                              17619.47 |      1.04 |
| NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | llama 8B Q4_0         | pp8096 |     10649.90 |                              10982.65 |      1.03 |
| NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | nemotron_h 9B BF16    | pp8096 |      7717.73 |                               7716.22 |      1.00 |
| NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | nemotron_h 9B Q4_K_M  | pp8096 |      7301.90 |                               7370.38 |      1.01 |
This commit is contained in:
Oliver Simons
2026-02-03 11:33:14 +01:00
committed by GitHub
parent e9a859db3c
commit 1f1e57f2bf
1 changed files with 13 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/mmq.cuh
+13 -6
View File
@@ -3697,13 +3697,20 @@ static __global__ void mul_mat_q(
tile_x_max_i, tile_y_max_j, kb0_start, kb0_stop);
}
template <ggml_type type, int mmq_x, bool need_check>
static __global__ void mul_mat_q_stream_k_fixup(
const int32_t * ids_dst, const int32_t * expert_bounds, float * __restrict__ dst, const float * __restrict__ tmp_last_tile,
const int ncols_x, const int nrows_x, const int ncols_dst, const int stride_col_dst,
const int nchannels_y, const int stride_channel_dst, const int nsamples_y, const int stride_sample_dst,
const int ncols_max) {
static __global__ void mul_mat_q_stream_k_fixup(const int32_t * ids_dst,
const int32_t * expert_bounds,
float * __restrict__ dst,
const float * __restrict__ tmp_last_tile,
const int ncols_x,
const int nrows_x,
const int ncols_dst,
const size_t stride_col_dst,
const int nchannels_y,
const size_t stride_channel_dst,
const int nsamples_y,
const size_t stride_sample_dst,
const int ncols_max) {
constexpr int mmq_y = get_mmq_y_device();
constexpr int qk = ggml_cuda_type_traits<type>::qk;
constexpr int ITER_K = get_iter_k(type);