fix(miopen): Build break with GCC 15. by benrichard-amd · Pull Request #9642 · ROCm/rocm-libraries

benrichard-amd · 2026-07-21T18:22:54Z

Motivation

Fix MIOpen build failure when compiling with clang++ against GCC 15's libstdc++ under C++20 mode. GCC 15's <ciso646> header emits a #warning that it is no longer a standard header since C++20, which becomes a fatal error due to MIOpen's -Werror flag.

Technical Details

Remove #include <ciso646> from projects/miopen/src/include/miopen/serializable.hpp.

<ciso646> defines alternative operator tokens (and, or, not, etc.) for C, but in C++ these are already language keywords — the include is a complete no-op. It was deprecated in C++17 and removed from the C++ standard in C++20.

The error surfaces when ROCm's clang++ uses GCC 15's libstdc++ as the host standard library since GCC 15 added a #warning to the header that fires under -std=c++20.

Test Plan

Build-only verification — the change is a no-op include removal with no effect on compiled output. CI tests confirm no regressions.

Test Result

MIOpen compiles successfully with the include removed.

Submission Checklist

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

<ciso646> is a no-op in C++ (the alternative operator tokens are already keywords) and was removed from the standard in C++20. GCC 15 libstdc++ emits a #warning when it is included under -std=c++20, which becomes fatal with -Werror. Co-Authored-By: Claude <noreply@anthropic.com>

therock-pr-bot · 2026-07-21T18:27:24Z

✅ All Checks Passed — Ready for Review

Check	Status	Details
📝 PR Description	✅ Pass	—
⛔ Forbidden Files	✅ Pass	—
🧪 Unit Test	⚠️ Warning	Error: Source/code files changed without an accompanying unit test. Expected: add at least one test file named like `test_<name>.py` / `test_<name>.cpp` (or `<name>_test.`). Current:* code file(s) changed: `projects/miopen/src/include/miopen/serializable.hpp`; no test file found
🔎 pre-commit	✅ Pass	—
🚫 Draft PR	🔜 To Be Enabled	—
🚩 Feature Flag	🔜 To Be Enabled	—
📊 Code Coverage	🔜 To Be Enabled	—
🤖 therock-pr-bot	✅ Pass	—

🎉 All checks passed! This PR is ready for review.

📖 Need help? See the Policy FAQ for details on every check and how to fix failures.

🙋 Wish to Override Policy?

therock-pr-bot · 2026-07-21T18:27:26Z

🎉 All checks passed! This PR is ready for review.

benrichard-amd · 2026-07-22T18:20:29Z

Hi @rocm/miopen-reviewers Can someone review this very trivial change? Thanks!

BrianHarrisonAMD

LGTM.

AmosLewis · 2026-08-10T21:30:42Z

Please fix the previous PR check:
Error: PR description must reference a JIRA ID, ISSUE ID, or a GitHub closing keyword.

one red is left, and it should be re-run before I bypass anything.

Check	Keypoint
`Windows gfx1151 / Test miopen (shard 3 of 4)`	Only red worth looking at. gtest aborted (`0x80000003`) in `ConvWinoFuryRxSf2x3Find2Fused/102`; shards 1/2/4 + `miopenprovider` passed, and this same shard passed Aug 7 on this same head `e4935b3` → not your diff. Re-run just that job
`Linux gfx94X / Configure test matrix`	Queued ~6h — repo-wide backlog, so the run reads `queued`, not running
`therock-pr-bot`	No `JIRA ID` / `ISSUE ID` line in the description; only you can add it

If that shard goes green you won't need me: mergeStateStatus is BLOCKED, not UNSTABLE, so your
auto-merge fires by itself. I'll --admin only if every remaining red is known infra.

ammallya · 2026-08-12T19:40:38Z

CI passes, flaky windows test machine

## Summary Backfills the **MIOpen 3.6.1 / ROCm 10.1.0** changelog section with externally-visible changes that landed since the `release/therock-10.0` branch but weren't yet documented. All 43 MIOpen commits in that range were reviewed; CI/build-infra, test-only, docs, version-bump, and already-documented changes were excluded. 21 entries added: ### Added * [RNN] Fused single-layer LSTM forward-inference kernel, gated behind `MIOPEN_DEBUG_RNN_FUSED_INFERENCE` (#8915) * [Conv] Naive reference kernel enabling 3D INT8 forward convolution (#10051) * [Conv] `ConvHipConv` solver backed by vendored hipconv kernels (#9746) * [Conv] float8 vector type support + deduplicated math kernels (#10130) * [Conv] Validated gfx1151 and gfx1200 (Navi) SystemDBs (#10050) * [Conv] `MIOPEN_DEBUG_DISABLE_SYSTEM_DB` / `MIOPEN_DEBUG_DISABLE_USER_DB` env vars (#10728) * [Conv] CK depthwise convolution on RDNA wave32 GPUs (#9373) ### Changed * [Conv] Winograd Rage RxS f2x3 on gfx950 + gfx942 v4_6_1/v4_9_1 kernels (#9014) * [Conv] Updated 3D AI heuristics for gfx942 & gfx950 (#8954) ### Optimized * [Conv] Single GEMM for point (1x1) output convolution, all directions, 2D/3D (#10912) * [Conv] Naive backward-data kernel work group widened to 1024 (#10684) ### Resolved Issues * [Conv] Batch-split miscalculation on large WrW tensors — ROCM-26011 (#11188) * [Conv] Conv failure on XNACK-unsupported gfx9 APUs (gfx90c) (#11211) * [BatchNorm] Device heap-overflow in fwd spatial `FinalMeanVariance` — ALMIOPEN-2424 (#11093) * [Conv] Deterministic `SetNextValue` polarity for grouped Bwd/Wrw CK solvers (#10240) * [LRN][LayerNorm] Grid alignment + OOB threads on gfx1250 (#10251) * [Conv] Stale gfx950 grouped-wrw SystemDB entries (#11130) * [Conv][Windows] System-DB path resolution from the MIOpen module (#10838) * [Conv] System-DB fallback so MI308X (gfx942) resolves a usable DB (#8802) * JSON perf logs (`MIOPEN_PERFORMANCE_LOGS`) dropping the last solver during Find (#11166) * Build failure against GCC 15 libstdc++ under C++20 (#9642) ## Notes Deliberately excluded: internal/default-off changes (#10219, #10088, #10649, #10231), a near-duplicate of an existing 10.0.0 entry (#10188), and CI/build-infra/test-only/docs commits. Docs-only change; no code impact. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

benrichard-amd requested a review from a team as a code owner July 21, 2026 18:22

github-actions Bot added the project: miopen label Jul 21, 2026

therock-pr-bot Bot added the Not ready to Review label Jul 21, 2026

assistant-librarian Bot added the organization: ROCm label Jul 21, 2026

BrianHarrisonAMD approved these changes Jul 22, 2026

View reviewed changes

benrichard-amd added 3 commits July 23, 2026 12:13

Merge branch 'develop' into users/benrichard-amd/miopen-gcc-15

6d53e8d

Merge branch 'develop' into users/benrichard-amd/miopen-gcc-15

3984bb4

Merge branch 'develop' into users/benrichard-amd/miopen-gcc-15

936f85a

benrichard-amd removed the Not ready to Review label Jul 31, 2026

therock-pr-bot Bot added the Not ready to Review label Jul 31, 2026

benrichard-amd removed the Not ready to Review label Jul 31, 2026

Merge branch 'develop' into users/benrichard-amd/miopen-gcc-15

e4935b3

therock-pr-bot Bot added the Not ready to Review label Aug 7, 2026

benrichard-amd enabled auto-merge (squash) August 8, 2026 02:19

benrichard-amd removed the Not ready to Review label Aug 10, 2026

Merge branch 'develop' into users/benrichard-amd/miopen-gcc-15

0581d4d

ammallya disabled auto-merge August 12, 2026 19:40

ammallya merged commit 2a2cb25 into develop Aug 12, 2026
155 of 169 checks passed

ammallya deleted the users/benrichard-amd/miopen-gcc-15 branch August 12, 2026 19:40

JonathanLichtnerAMD mentioned this pull request Sep 1, 2026

docs(miopen): backfill missing ROCm 10.1.0 changelog entries #11586

Merged

benrichard-amd mentioned this pull request Sep 3, 2026

[issue] Build failure: GCC 15 / libstdc++-15: MIOpen includes <ciso646>, removed in C++20 #9352

Closed

lucbruni-amd mentioned this pull request Sep 8, 2026

fix(miopen): remove obsolete <ciso646> include (GCC 15 build failure) #9718

Closed

1 task done

Navigation Menu

Repository navigation

fix(miopen): Build break with GCC 15. - #9642

fix(miopen): Build break with GCC 15.#9642
ammallya merged 6 commits into
developfrom
users/benrichard-amd/miopen-gcc-15

benrichard-amd commented Jul 21, 2026 •

edited

Loading

Uh oh!

therock-pr-bot Bot commented Jul 21, 2026 •

edited

Loading

Uh oh!

therock-pr-bot Bot commented Jul 21, 2026 •

edited

Loading

Uh oh!

benrichard-amd commented Jul 22, 2026

Uh oh!

BrianHarrisonAMD left a comment

Uh oh!

AmosLewis commented Aug 10, 2026 •

edited

Loading

Uh oh!

Uh oh!

ammallya commented Aug 12, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

benrichard-amd commented Jul 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Motivation

Technical Details

Test Plan

Test Result

Submission Checklist

Uh oh!

therock-pr-bot Bot commented Jul 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅ All Checks Passed — Ready for Review

Uh oh!

therock-pr-bot Bot commented Jul 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

benrichard-amd commented Jul 22, 2026

Uh oh!

BrianHarrisonAMD left a comment

Choose a reason for hiding this comment

Uh oh!

AmosLewis commented Aug 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

ammallya commented Aug 12, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

benrichard-amd commented Jul 21, 2026 •

edited

Loading

therock-pr-bot Bot commented Jul 21, 2026 •

edited

Loading

therock-pr-bot Bot commented Jul 21, 2026 •

edited

Loading

AmosLewis commented Aug 10, 2026 •

edited

Loading