Skip to content

Actions: AMD-Ecosystem/llama.cpp

Actions

Code Style Checker

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
code-style.yml
190 workflow runs
190 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #123: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #121: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #120: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
mmq: compacted MoE tiling for RDNA3.5 (default on) Code Style Checker #115: Pull request #63 synchronize by roberteg16
14s rogarcia.mmq-moe-compact-tiling
rogarcia.mmq-moe-compact-tiling
14s
mmq: compacted MoE tiling for RDNA3.5 (default on) Code Style Checker #114: Pull request #63 synchronize by roberteg16
55s rogarcia.mmq-moe-compact-tiling
rogarcia.mmq-moe-compact-tiling
55s
mmq: compacted MoE tiling for RDNA3.5 (default on) Code Style Checker #113: Pull request #63 synchronize by roberteg16
43s rogarcia.mmq-moe-compact-tiling
rogarcia.mmq-moe-compact-tiling
43s
mmq: compacted MoE tiling for RDNA3.5 (default on) Code Style Checker #112: Pull request #63 opened by roberteg16
26s rogarcia.mmq-moe-compact-tiling
rogarcia.mmq-moe-compact-tiling
26s
RDNA3.5 (gfx11 / gfx1151) MMQ prefill optimizations Code Style Checker #111: Pull request #32 synchronize by liangliangchang
35s liangliangchang:lichang.mmq_opt
liangliangchang:lichang.mmq_opt
35s
RDNA3.5 (gfx11 / gfx1151) MMQ prefill optimizations Code Style Checker #110: Pull request #32 synchronize by liangliangchang
25s liangliangchang:lichang.mmq_opt
liangliangchang:lichang.mmq_opt
25s
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #109: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #108: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
RDNA3.5 (gfx11 / gfx1151) MMQ prefill optimizations Code Style Checker #107: Pull request #32 synchronize by liangliangchang
17s liangliangchang:lichang.mmq_opt
liangliangchang:lichang.mmq_opt
17s
Merge pull request #56 from ROCm/rogarcia.enable-by-default-ggml_cuda… Code Style Checker #106: Commit 6e9f948 pushed by mgehre-amd
56s gfx11
gfx11
56s
tests: add MoE MMQ benchmark with routing-distribution generator Code Style Checker #105: Pull request #62 opened by roberteg16
1d 0h 0m 1s rogarcia.moe-mmq-benchmark
rogarcia.moe-mmq-benchmark
1d 0h 0m 1s
ggml-cuda: add dequant-float matvec (mmvdq) for Q4_K/Q5_K/Q6_K on RDN… Code Style Checker #104: Pull request #61 opened by Annieren
20s annier.mmvq-to-mmvdq
annier.mmvq-to-mmvdq
20s
Gfx11 Code Style Checker #103: Pull request #60 opened by Annieren
14s gfx11
gfx11
14s
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #102: Pull request #59 synchronize by jeffli-xilinx
24s jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
24s
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #101: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #100: Pull request #59 synchronize by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
gfx1151 (Strix Halo): fuse attn_k+v into single MMVQ dispatch Code Style Checker #99: Pull request #59 opened by jeffli-xilinx
Action required jeffli-xilinx:rdna3_5-kv-fusion
jeffli-xilinx:rdna3_5-kv-fusion
Action required
ggml-cuda: GEMM weight row padding + one-time K-padded f16 dequant for prefill Code Style Checker #97: Pull request #57 synchronize by roberteg16
1m 4s rogarcia.cuda-gemm-pad-prefill-dequant
rogarcia.cuda-gemm-pad-prefill-dequant
1m 4s
ggml-cuda: GEMM weight row padding + one-time K-padded f16 dequant for prefill Code Style Checker #96: Pull request #57 opened by roberteg16
32s rogarcia.cuda-gemm-pad-prefill-dequant
rogarcia.cuda-gemm-pad-prefill-dequant
32s
ggml-cuda: enable GGML_CUDA_GRAPH_OPT by default on RDNA3.5 Code Style Checker #95: Pull request #56 opened by roberteg16
46s rogarcia.enable-by-default-ggml_cuda_graph_opt-on
rogarcia.enable-by-default-ggml_cuda_graph_opt-on
46s
Merge pull request #53 from ROCm/rogarcia.roofline-moe-tokens-per-expert Code Style Checker #94: Commit 0941c4c pushed by mgehre-amd
42s gfx11
gfx11
42s
CUDA: fused chunked gated_delta_net kernel (RDNA3.5) Code Style Checker #91: Pull request #54 synchronize by roberteg16
22s rogarcia.gdn-shared-kq
rogarcia.gdn-shared-kq
22s