Repository navigation
feat(miopen): add gfx1250 build to the MIOpen dev image - #12496
Merged
JonathanLichtnerAMD merged 1 commit intoSep 29, 2026
Merged
Conversation
Add gfx1250 and gfx1250-strict to the arch lists that produce the published rocm/miopen-dev image (getDockerImage's "ci" branch) and the :therock base image (buildTheRockDockerImage), so CK and MIOpen build kernels for gfx1250 hardware. Verified in the current dev image toolchain that both offload-archs compile cleanly and produce distinct device codeobjects. JIRA: ALMIOPEN-2644
✅ All Checks Passed — Ready for Review
📖 Need help? See the Policy FAQ for details on every check and how to fix failures. |
|
🎉 All checks passed! This PR is ready for review. |
averinevg
approved these changes
Sep 25, 2026
JonathanLichtnerAMD
deleted the
users/jlichtne/almiopen-2644-gfx1250-dev-image
branch
September 29, 2026 16:22
Contributor
TheRock Submodule Bump ActivityNewest first
|
JonathanLichtnerAMD
added a commit
that referenced
this pull request
Oct 2, 2026
…12800) JIRA ID: ALMIOPEN-2820 ## Motivation Reverts #12496 (320f7a8), which added `gfx1250;gfx1250-strict` to the MIOpen dev image arch lists. With those archs, the CK stage of the Docker image build (`build_ck`, Dockerfile:375) fails: ``` ld.lld: error: local memory (131552) exceeds limit (65536) ``` CK's `CK_USE_GFX1250`-gated large-LDS wmma grouped-conv instances are also compiled for gfx1101/gfx1151 (64 KB LDS). This broke the MIOpen develop nightly (Jenkins `rocm-libraries-folder/MIOpen/develop` #1589 and #1590) at "Build Docker for CI"; the last good run was #1588. The break was masked for about a day because the CI image tag hash does not include the arch list, so #1584-#1588 reused a stale image until an unrelated CK `CMakeLists.txt` change (#12686) forced a rebuild. The CK-side fix is tracked in AICK-2468. gfx1250 can be re-added to the dev image once that lands. ## Technical Details In `projects/miopen/vars/utils.groovy`, remove `gfx1250;gfx1250-strict` from two arch lists: - the `ci` list in `getDockerImage()` - the `:therock` multiarch list in `buildTheRockDockerImage()` ## Risk Assessment **Risk Level: 🟢 Low** ### Impacted Components MIOpen Jenkins CI Docker image build (`projects/miopen/vars/utils.groovy`) only; no library code changes. ### Potential Side Effects gfx1250 is temporarily absent from the `rocm/miopen-dev` image. ### Mitigation Steps Re-add gfx1250 to the dev image once the CK fix (AICK-2468) lands. ## Test Plan Verify on the next develop nightly: the forced CI image rebuild completes `build_ck`, then Build Docker for CI, Full Tests, Publish Dev Image and Nightly Tests pass. ## Test Result Not yet run; verification is the next develop nightly (see Test Plan).
mattwill-amd
pushed a commit
that referenced
this pull request
Oct 14, 2026
JIRA ID: ALMIOPEN-2913 ## Motivation Only 5 of the 30 commits that landed against `projects/miopen` since the 3.6.2 version bump (`2d20a8f775e`, #12211) were reflected in `CHANGELOG.md`. The remaining 25 were reviewed and the 15 externally-visible ones documented, so the ROCm 10.2.0 section accurately reflects what shipped. 14 entries added (two of the fixes share one entry), plus a duplicate-heading fix. ### Added * [Conv] NHWC (channel-last) support in the GEMM convolution solvers, with new im2col/col2im kernels (#8693) * [Conv] Cross-architecture AI heuristics for immediate mode (`MIOPEN_FIND_MODE=2`), extending data-driven selection to Navi (#9680) * [Conv] Missing 3x3 stride-1 pad-1 tile-28 CK depthwise backward-data instance — ALMIOPEN-2423 (#11271) ### Changed * [Conv] `ConvHipConv` held out of the AI perf-config picker until its model is retrained — ALMIOPEN-2494 (#13337) * [Conv] CK byte-span gate lifted for forward and backward-data, now that CK fixes the 32-bit offset overflow itself — ALMIOPEN-2645 (#12396) ### Optimized * [Conv] `ConvDirectNaiveConv`: per-direction kernel files, float accumulators, spatial tiling, bfloat16 path — ALMIOPEN-2485/2486 (#7529) * Kernel cache reuses cached SQLite connections instead of reopening per load/store — ALMIOPEN-2598 (#11933) ### Resolved Issues * [Conv] Incorrect backward-weights results from the gfx950 direct kernel at `waves_q == 1` (missing cross-lane ordering on LDS staging reuse) — ROCM-31508 (#12406) * [Conv] Up to 8x slower hipconv `direct`/`direct_l1`/`direct_wgrad` under ROCm 10.1's full AMDGCN LTO default; fixed with `-fno-offload-lto` — ROCM-31334 (#12328) * [Conv] gfx942 multi-node training throughput regression from backward GEMM on small-batch 1x1 NHWC shapes — ROCM-31155 (#13232) * [Conv] Invalid CK perf-config entries in the gfx1201 and gfx1100 system databases — ROCM-29026 (#12088, #12117) * [BatchNorm] JIT compilation failures on non-GFX9 targets; Welford DPP assembly restricted to GFX9 — ALMIOPEN-2838 (#12889) * Out-of-bounds write in `miopenLoadSolution` when a serialized solution carries more than three work dimensions — ROCM-31082 (#12242) * `uint64_t` typedef conflict between `miopen_cstdint.hpp` and system `<stdint.h>` on LP64 Linux (#12623) ## Technical Details Single-file change to `projects/miopen/CHANGELOG.md`. The 10.2.0 section had **two** `### Resolved Issues` headings. The two entries under the trailing one (`ConvHipConv` gfx1250 crash/LDS race, and the perf-db reload config loss) were merged into the first, and the trailing heading removed — this accounts for the 2 removed bullets in the diff. Section order is now `Resolved Issues → Added → Changed → Optimized → Removed`, matching the 3.6.1 section. `Optimized` is new to this release section. Already documented, so untouched: #12874 (reduce int32 rejection), #12000 / #13016 / #13213 (hipconv v0.3.1 / v0.3.2 / v0.4.0), #11761 (gfx803 removal). ## Notes Deliberately excluded as not externally visible (10 commits): * **hipDNN convolution forwarding (#12276)** — `MIOPEN_ENABLE_HIPDNN_WRAPPER` defaults to `OFF` (`projects/miopen/CMakeLists.txt:162`), so no shipped build compiles the forwarding code and the runtime `MIOPEN_HIPDNN_FORWARDING` flag has nothing to enable. * **Test-only**: #13289 (gfx1250 depthwise test applicability, ALMIOPEN-2841), #11610 (hipDNN parity harness), #12660 (Jenkins status post). * **Internal kernel-math cleanup, generated ISA unchanged**: #13312 (drop direct `__ocml` calls), #13190 (remove broken single-argument `__half` overloads). * **Add/revert pairs, net no change**: #12466 + #13066 (gfx1101 Navi system DBs), #12496 + #12800 (gfx1250 in the dev image arch lists). ## Risk Assessment Documentation-only change to a single Markdown file; no code, build, or runtime impact. **Risk Level: 🟢 Low** ## Test Plan None required — Markdown-only. Verified that the entry topics are not already documented elsewhere in the file, that no content was lost when merging the duplicate heading, and that every new entry maps to a commit in `2d20a8f775e..origin/develop -- projects/miopen`. ## Test Result 16 bullets added, 2 relocated — 14 net-new entries covering 15 commits. Duplicate heading removed; section ordering matches 3.6.1. Co-authored-by: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
feat(miopen): add gfx1250 build to the MIOpen dev image
JIRA ID: ALMIOPEN-2644
Description
Adds
gfx1250andgfx1250-strictto the two GPU-arch lists inprojects/miopen/vars/utils.groovythat drive MIOpen's dev Docker images:getDockerImage()'scibranch (the arch list published asrocm/miopen-devviapublishDevDockerImage()) andbuildTheRockDockerImage()'s:therockbase-image arch list. These listsare threaded automatically into the Dockerfile via
--build-arg THEROCK_ASIC, which flows into CK's-D GPU_TARGETSand rbuild's-DGPU_ARCHS, so no other files need changes. gfx1250/gfx1250-strict werealready supported as build targets elsewhere (CK, MIOpen's cdna5 arch libs,
hiptensor, rocWMMA); this closes the gap in the two dev-image arch lists
that had been missed.
Risk Assessment
Risk Level: Low
Two-line, additive-only change to a Jenkins/CI build-config list (no
existing archs removed or reordered). Verified inside a
rocm/miopen-dev:latestcontainer that a fresh MIOpen CMake configure with-DGPU_TARGETS="gfx1250;gfx1250-strict"succeeds, and thatlibMIOpenCKGroupedConv_gfx1250-strict.sobuilds with a real ~318MB.hip_fatbinsection containing compiled device kernels (confirmed viaclang-offload-bundler/objcopy, not just symbol-name acceptance).Test Result
Build-only enablement — no gfx1250 hardware exists yet in TheRock's CI
matrix, so there is no on-hardware test to run. Validated via the manual
CMake configure/build described above.