[XiaobingSuper]

What

Adds the 62 missing N=3072, K=128 rows to a8w8_bpreshuffle_tuned_gemm_kimik3.csv, plus the matching shapes in the untuned CSV.

Why

Kimi-K3's KDA gate projection f_b_proj is [M, 128] x [128, 128 * heads/tp]. K is the KDA head dim, 128 at any tp; N is num_local_heads * 128, so the 96 KDA heads give N=1536 at tp8 and N=3072 at tp4. Only 1536 was tuned, so tp4 misses the table on every token.

A miss here is not a slower kernel. gemm_a8w8_bpreshuffle falls through to gemm_a8w8_bpreshuffle_ck, whose rowwise_heuristic_dispatch has branches only for K >= 512 and K >= 192 && K % 64 == 0. K=128 hits:

TORCH_CHECK(false, "Unsupported K for heuristic dispatch: 128. Supported K greater than 192 and K % 64 == 0.")

so the process raises on its first token. Tuning confirms there is no CK kernel for this shape at all: every ck candidate reported This GEMM is not supported!, and every winner is cktile or flydsl.

How

Same M grid the N=1536 rows carry (64-step between 1024 and 4096), so padded-M lookups behave the same at both widths.

python3 csrc/ck_gemm_a8w8_bpreshuffle/gemm_a8w8_bpreshuffle_tune.py \
  -i aiter/configs/model_configs/a8w8_bpreshuffle_untuned_gemm_kimik3.csv \
  -o aiter/configs/model_configs/a8w8_bpreshuffle_tuned_gemm_kimik3.csv \
  --mp 8 --errRatio 0.01

62 shapes, zero failures. 5 pick cktile at small M, 57 flydsl above -- the same split N=1536 has.

The rows are inserted in the file's existing (N, M, K) order rather than letting the tuner re-sort the whole file, so the diff is 62 additions and no deletions.

Verification

gfx950 / 256 CU, via gemm_a8w8_bpreshuffle directly:

shape	before	after
tp8 M=1..16384 N=1536 K=128	OK	OK
tp4 M=1..16384 N=3072 K=128	FAIL Unsupported K for heuristic dispatch: 128	OK

End-to-end: Kimi-K3 at --pipeline-parallel-size 2 --tensor-parallel-size 4 on MI355X x8 now serves with no framework-side GEMM override, GSM8K 5-shot over the full 1319 questions at 0.9598 flexible-extract / 0.9606 strict-match, inside the range recipes/Kimi-K3.md records for this model (0.9591-0.9659 / 0.9591-0.9666).

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[github-actions]

🏷️ CI Guide

Runs automatically on every PR:

• ✅ Pre-checks (submodule verification, code formatting)
• ✅ Aiter op tests (gfx942 + gfx950)
• ✅ Triton tests on MI35X (only when aiter/ops/triton/** or related paths are changed)

Extended tests (opt-in via labels):

Label	Tests
ci:gfx1250-ffm-triton	Run the five-shard gfx1250 FFM Triton test suite
ci:triton-300x	Run an additional Triton test job on MI300X in PRs; main branch always runs both MI35X and MI300X
multigpu	Aiter multi-GPU tests on the 8-GPU runner
ci:sglang	SGLang integration tests: DeepSeek-R1-MXFP4 accuracy, Qwen 3.5 accuracy
ci:atom	ATOM benchmark: DeepSeek-R1-0528, GPT-OSS-120B
ci:atom_full	ATOM accuracy suite for PR and main models from ATOM models_accuracy.json
ci:vllm	vLLM benchmark: GPT-OSS-120B, DeepSeek-R1-0528, Kimi-K2.5
ci:all	All standard extended tests (excludes ci:atom_full)

Only add ci:atom_full for FlyDSL or Triton upgrades.
Add labels via the sidebar or gh pr edit 5783 --add-label <label>

PR title tags & labels:
Component tags ([Triton/Gluon], [HIP], [CK], [ASM], ...) are added to the PR title and as PR labels automatically from the changed files and re-synced on every push — change-type tags like [fix]/[Perf], op tags like [MLA], and human labels (ci:*) are left untouched. Add the no-auto-title label to opt this PR out.

[yifehuan]

lgtm