Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HIP: add fattn-mma-f16 for RDNA4 (#18481)

Browse Source 
* finish VQ mma

* flash_attn_ext_f16_iter

* KQ_rowsum

* correct exp

* fix scale error

* fix softmax scale

* fix softmax scale

* enable fattn on cpu side

* fix random error

* disable fattn-mma-f16 on rdna3

* fix wrong col for rdna

* use identity mat to transpose

* resolve conflicts

* basic tuning for DeepSeek-R1-Distill-Qwen-1.5B

* fix volta compile error

* align rdna4 policy for fattn

* adjust fattn policy

* adjust kernel selection logic

* update as the review comments

* keep fattn-wmma logic

* adjust kernel selection logic

---------

Co-authored-by: zhang hui <you@example.com>
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
This commit is contained in:
yulo
2026-01-13 13:52:16 +01:00
committed by GitHub
co-authored by zhang hui Johannes Gäßler
parent c1e79e610f
commit ea4a321f2a
6 changed files with 266 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/fattn-common.cuh
+1 -1
View File
@@ -914,7 +914,7 @@ void launch_fattn(
const int nblocks_stream_k = max_blocks;
const bool use_stream_k = cc >= GGML_CUDA_CC_ADA_LOVELACE || tiles_efficiency_percent < 75;
const bool use_stream_k = cc >= GGML_CUDA_CC_ADA_LOVELACE || amd_wmma_available(cc) || tiles_efficiency_percent < 75;
blocks_num.x = use_stream_k ? nblocks_stream_k : ntiles_total;
blocks_num.y = 1;