[rahulbatra85]

Motivation

Add tuned GEMM configs for Inkling Model Shapes

Technical Details

Adds JSON files

Test Plan

Unit test

Test Result

All tests should pass

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; 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 5922 --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.

[Copilot]

Copilot review overview

🔵 Needs a closer look

The hardware-specific runtime and performance behavior requires head-matched gfx950 validation.

Review effort: Balanced
Findings: None

What changed in this PR

Adds gfx950 Triton BF16 GEMM tuning for two Inkling model shapes.

Changes:

• Adds M-bucketed configs for (N=3328, K=6144) and (N=2816, K=6144).
• Records M=64 benchmark results in the Inkling model configuration.

File	Description
aiter/​ops/​triton/​configs/​gfx950/​triton/​gemm/​gemm_a16w16/​GEMM-A16W16-N=3328-K=6144.json	Adds the 3328×6144 tuning table.
aiter/​ops/​triton/​configs/​gfx950/​triton/​gemm/​gemm_a16w16/​GEMM-A16W16-N=2816-K=6144.json	Adds the 2816×6144 tuning table.
aiter/​configs/​model_configs/​inkling_bf16_tuned_gemm.csv	Records Inkling benchmark results.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[azaidy]

LGTM!