[AIROCMLIR-1061] Increase rocm-smi timeout

[justinrosner]

Motivation

This PR addresses some flaky/intermittent behavior observed in the tuningRunner-* LIT tests when calling into rocm-smi for GPU topology discovery.

Technical Details

In https://github.com/ROCm/rocmlirTriton/pull/310 in rocmlirTriton we bumped the timeout from 10s -> 60s. Looking at the history of nightly CI runs since then shows that this issue has not come up. This PR just ports those same changes to rocMLIR.

Test Plan

• Run the nightly CI

Test Result

• Nightly CI

Submission Checklist

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

[Copilot]

Pull request overview

This PR reduces flakiness in tuningRunner-* LIT tests by increasing the rocm-smi GPU topology discovery timeout, aligning rocMLIR with the previously proven change in rocmlirTriton.

Changes:

• Increase rocm-smi invocation timeout from 10s to 60s during GPU topology discovery.
• Add an inline comment explaining why a larger timeout is necessary on large multi-GPU systems.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[rocmlir-pr-reviewer]

Verdict: APPROVE -- submitted as COMMENT (automated reviews are advisory)  ·  Findings: 0 (0 Critical, 0 Major, 0 Minor)

---

Scope

Bumps the rocm-smi subprocess timeout in GpuTopology.discover() from 10s to 60s and adds an explanatory comment, addressing flaky tuningRunner-* LIT tests. Single-file, single-line functional change (mlir/utils/performance/tuningRunner.py), porting a fix already validated in rocmlirTriton.

Findings

No blocking issues found.

Notes

• The added comment cites ~20s enumeration time while the timeout is set to 60s; the extra headroom is reasonable given the flakiness motivation, so no change requested.
• Change is a pure constant bump on an existing, tested code path; no new op/pass/API surface, so no additional Lit/E2E coverage is warranted.

CI status

No non-self checks report fail or cancel. Python format/lint and Python performance script tests pass; C/C++ premerge is still in progress at review time.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #2428      +/-   ##
===========================================
+ Coverage    82.57%   82.86%   +0.30%     
===========================================
  Files          120      120              
  Lines        42852    42828      -24     
  Branches      7110     7106       -4     
===========================================
+ Hits         35381    35489     +108     
+ Misses        4815     4696     -119     
+ Partials      2656     2643      -13     

Flag	Coverage Δ
gfx120x	82.74% <ø> (+0.22%)	⬆️
gfx950	82.68% <ø> (+0.34%)	⬆️
mfma	82.63% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.
see 29 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.