[truong-v]

Motivation

Fixes #6317

The top-k selection in topk_gating_kernel and topk_gating_kernel_smem_n scans only full vectors, so with a vec_size that does not divide num_experts the last experts are never selected. Dispatch never does this today, so this PR makes the precondition explicit and fails loudly instead of misrouting.

Technical Details

• In both kernels, if(num_experts % vec_size != 0) __builtin_trap(); right after num_experts_vec is computed.
• Adding a scalar tail to the selection loops instead made topk_gating_kernel 2.6-5.8% slower at T <= 1024 for divisible num_experts, so this PR uses the guard.

No test is added, because the condition cannot be produced through aiter.ops.topk.topk_gating.

Test Plan

python op_tests/test_moe_topk_gating.py

Test Result

MI325X (gfx942): all 4560 configurations pass. In a standalone harness with a forced vec_size, all 6 non-divisible cases now trap instead of returning wrong ids. Kernel time is 0.988-1.025x of base (median 1.001x).

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/TheRock/blob/main/GOVERNANCE.md#pull-requests.

[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 (added automatically when gfx942 configs change); main branch always runs both MI35X and MI300X
ci:triton-355	Run the full Triton test suite on MI35X, not only the tests the change affects
ci:multi-gpu	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 6323 --add-label <label>

One backend per PR:
A PR changes one kernel backend: [Triton/Gluon] (Triton and Gluon count as one), [HIP], [ASM], [CK], [OPUS] or [FlyDSL]. If the title ends up with two backend tags, split the PR -- as stacked pull requests when one part cannot merge without the other.

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 stop the title rewrites; labels stay in sync either way.