Skip to content

[AIMIGRAPHX-1186] TheRock packaging: amdrocm-migraphx naming + drop weak deps (per-arch packaging in UFB)#5086

Open
kentqian wants to merge 5 commits into
developfrom
kqian1/meta-package-fix
Open

[AIMIGRAPHX-1186] TheRock packaging: amdrocm-migraphx naming + drop weak deps (per-arch packaging in UFB)#5086
kentqian wants to merge 5 commits into
developfrom
kqian1/meta-package-fix

Conversation

@kentqian

@kentqian kentqian commented Jul 22, 2026

Copy link
Copy Markdown
Contributor

Summary

  • Rename the package to amdrocm-migraphx per RFC0009 (Package name now matches the .deb/.rpm filename and the amdrocm-* ecosystem).
  • Drop the weak Recommends/Suggests used for per-GPU ROCm libs.

Technical Details

One fat compile with multi-archs; UFB re-runs cpack once per GPU arch → amdrocm-migraphx-gfxNNN, each hard-depending on that arch's amdrocm-dnn/blas-gfxNNN + amdrocm-runtime.

Compatibility / scope

  • All changes inside TheRock; default packaging and migraphx's own CI unaffected.
  • migraphx's own device code stays a single fat ELF (one compile); only naming/deps change.

Arch-specific dependencies (moved out of CMake)

  • Per-GPU device-library dependencies are no longer declared in CMake. In the TheRock repo, MIOpen/rocBLAS ship only as per-arch packages (amdrocm-dnn-gfxNNN / amdrocm-blas-gfxNNN) with no arch-neutral variant, so CMake now declares only the arch-neutral amdrocm-runtime dependency.
  • The per-arch device deps are injected externally at packaging time (one fat build, repackaged per arch). To produce a package with device dependencies, they must be added externally per target arch (amdrocm-dnn-<arch>, amdrocm-blas-<arch>).
  • Encoding per-arch dependency selection inside CMake would be cumbersome and redundant, and no longer fits the original single-package ROCm layout — hence it is intentionally kept out of the migraphx build.

Changelog Category

Add a CHANGELOG.md entry for any option other than Not Applicable

    • Added: New functionality.
    • Changed: Changes to existing functionality.
    • Removed: Functionality or support that has been removed. (Compared to a previous release)
    • Optimized: Component performance that has been optimized or improved.
    • Resolved Issues: Known issues from a previous version that have been resolved.
    • Not Applicable: This PR is not to be included in the changelog.

Follow the LLVM AI Tool Use Policy for contributions using AI.

@kentqian kentqian self-assigned this Jul 22, 2026
@kentqian
kentqian requested a review from causten as a code owner July 22, 2026 18:53
Copilot AI review requested due to automatic review settings July 22, 2026 18:53
@kentqian kentqian added the dependencies Pull requests that update a dependency file label Jul 22, 2026
@kentqian kentqian changed the title [AIMIGRAPHX-1186] meta package fix [AIMIGRAPHX-1186] TheRock: deb/rpm per-arch meta packages for external ROCm deps (drop weak Recommends/Suggests) Jul 22, 2026

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This pull request updates the TheRock packaging path in CMakeLists.txt (MIGRAPHX_PACKAGE_BACKEND=therock) to shift per-GPU dependency selection to per-arch meta packages with hard dependencies, and aligns the TheRock package naming with the amdrocm-* ecosystem (RFC0009). The default (non-TheRock) ROCm packaging path is intended to remain functionally unchanged.

Changes:

  • Switch TheRock package naming from MIGraphX to amdrocm-migraphx and plumb the package name into rocm_create_package(NAME ...).
  • For per-arch mode, create per-arch meta package components (e.g., gfx90a) plus an all meta component, installing a small README into each component’s doc dir.
  • Retain the existing device-all dependency behavior when MIGRAPHX_THEROCK_GPU_ARCH is empty.

Comment thread CMakeLists.txt Outdated
Comment on lines +475 to +477
install(FILES ${_MGX_META_DOC}
DESTINATION share/doc/${_MGX_PKG_NAME}-${mgx_gpu_arch}
COMPONENT ${mgx_gpu_arch})
Comment thread CMakeLists.txt Outdated
Comment on lines +485 to +487
install(FILES ${_MGX_META_DOC}
DESTINATION share/doc/${_MGX_PKG_NAME}-all
COMPONENT all)
Comment thread CMakeLists.txt
Comment on lines +502 to +503
# Legacy ROCm packaging.
set(_MGX_PKG_NAME "MIGraphX")
@gh-app-migraphx-bot-pr-write

gh-app-migraphx-bot-pr-write Bot commented Jul 22, 2026

Copy link
Copy Markdown
Test Batch New Rate (52190c) Old Rate (8c5dbe)* Diff Status
torchvision-resnet50 64 3,206.49 3,266.06 -1.82%
torchvision-resnet50_fp16 64 7,533.11 7,544.96 -0.16%
torchvision-densenet121 32 2,486.73 2,482.83 0.16%
torchvision-densenet121_fp16 32 1,278.78 4,991.72 -74.38% 🔴
torchvision-inceptionv3 32 2,037.75 2,057.53 -0.96%
torchvision-inceptionv3_fp16 32 3,470.41 4,407.39 -21.26% 🔴
cadene-inceptionv4 16 802.19 819.62 -2.13%
cadene-resnext64x4 16 782.68 783.03 -0.05%
slim-mobilenet 64 8,422.02 8,376.48 0.54%
slim-nasnetalarge 64 189.10 229.56 -17.63% 🔴
slim-resnet50v2 64 3,178.80 3,181.03 -0.07%
bert-mrpc-onnx 8 nan 1,170.53 nan
bert-mrpc-tf 1 nan 490.19 nan
pytorch-examples-wlang-gru 1 473.08 479.41 -1.32%
pytorch-examples-wlang-lstm 1 374.42 420.54 -10.97% 🔴
torchvision-resnet50_1 1 1,033.19 1,040.50 -0.70%
cadene-dpn92_1 1 167.52 445.19 -62.37% 🔴
cadene-resnext101_1 1 365.29 365.17 0.03%
onnx-taau-downsample 1 nan 847.89 nan
dlrm-criteoterabyte 1 24.79 32.42 -23.54% 🔴
dlrm-criteoterabyte_fp16 1 52.56 51.76 1.54%
agentmodel 1 9,773.16 9,166.41 6.62% 🔆
unet_fp16 2 nan 58.78 nan
resnet50v1_fp16 1 1,440.51 1,359.75 5.94% 🔆
resnet50v1_int8 1 1,617.55 1,611.83 0.35%
bert_base_cased_fp16 64 nan 1,098.77 nan
bert_large_uncased_fp16 32 nan 345.61 nan
bert_large_fp16 1 nan 207.09 nan
distilgpt2_fp16 16 nan 2,096.49 nan
yolov5s 1 nan 559.61 nan
tinyllama 1 nan 45.79 nan
vicuna-fastchat 1 nan 44.16 nan
whisper-tiny-encoder 1 nan 412.52 nan
whisper-tiny-decoder 1 nan 408.98 nan
llama2_7b 1 nan 20.83 nan
qwen1.5-7b 1 nan 23.51 nan
phi3-3.8b 1 nan 26.70 nan
llama3-8b 1 nan 21.63 nan
whisper-large-encoder 1 nan 10.18 nan
whisper-large-decoder 1 nan 104.98 nan
mistral-7b 1 nan 23.63 nan
FLUX.1-schnell 1 nan 760.21 nan

Regressions detected 🔴

* No develop baseline was found for this PR's branch point; compared against the latest available develop run instead.

@gh-app-migraphx-bot-pr-write

gh-app-migraphx-bot-pr-write Bot commented Jul 22, 2026

Copy link
Copy Markdown
Test Status Result
bert-mrpc-onnx ERROR - check error output
traceback
2026-07-24 16:42:38.590838 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:42:38.590893 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134354201351050.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
bert-mrpc-tf ERROR - check error output
traceback
2026-07-24 16:42:45.276039 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:42:45.276108 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134360886558133.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
pytorch-examples-wlang-gru PASSED: MIGraphX meets tolerance
pytorch-examples-wlang-lstm PASSED: MIGraphX meets tolerance
dlrm-criteoterabyte PASSED: MIGraphX meets tolerance
agentmodel PASSED: MIGraphX meets tolerance
unet ERROR - check error output
traceback
2026-07-24 16:45:23.243238 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:45:23.243305 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134518853756308.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
resnet50v1 PASSED: MIGraphX meets tolerance
bert_base_cased_fp16 ERROR - check error output
traceback
2026-07-24 16:45:55.258012 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:45:55.258096 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134550868541203.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
bert_large_uncased_fp16 ERROR - check error output
traceback
2026-07-24 16:46:02.607789 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:46:02.607859 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134558218308909.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
bert_large ERROR - check error output
traceback
2026-07-24 16:46:05.907813 [WARN] [/data/src/onnx/onnx_parser.cpp:296] Model has unbound symbolic dimension(s): batch_size. These default to 1 and may cause unexpected behavior. Try setting --dim-param @<name> <value> or --input-dim @<input> <dims> if program compilation fails.
2026-07-24 16:46:12.987571 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:46:12.987632 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134568598083410.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
yolov5s ERROR - check error output
traceback
2026-07-24 16:46:19.815878 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:46:19.815936 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134575426394486.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
tinyllama ERROR - check error output
traceback
2026-07-24 16:46:37.466219 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:46:37.466286 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134593076740538.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
vicuna-fastchat ERROR - check error output
traceback
2026-07-24 16:47:01.595561 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:47:01.595646 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134617206086132.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
whisper-tiny-encoder ERROR - check error output
traceback
2026-07-24 16:47:15.969985 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:47:15.970045 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134631580499328.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
whisper-tiny-decoder ERROR - check error output
traceback
2026-07-24 16:47:19.440084 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:47:19.440148 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134635050600487.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
distilgpt2_fp16 ERROR - check error output
traceback
2026-07-24 16:47:22.722273 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:47:22.722342 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134638332793074.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
llama2_7b PASSED: MIGraphX meets tolerance
qwen1.5-7b ERROR - check error output
traceback
2026-07-24 16:50:26.045045 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:50:26.045112 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134821655563181.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
phi3-3.8b ERROR - check error output
traceback
2026-07-24 16:52:15.117898 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:52:15.117966 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution134930728416848.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
llama3-8b ERROR - check error output
traceback
2026-07-24 16:52:50.801210 [WARN] [/data/src/onnx/onnx_parser.cpp:296] Model has unbound symbolic dimension(s): batch_size, sequence_length. These default to 1 and may cause unexpected behavior. Try setting --dim-param @<name> <value> or --input-dim @<input> <dims> if program compilation fails.
2026-07-24 16:54:01.221456 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:54:01.221518 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution135036831967843.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
whisper-large-encoder ERROR - check error output
traceback
2026-07-24 16:55:25.585755 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:55:25.585823 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution135121196273448.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
whisper-large-decoder PASSED: MIGraphX meets tolerance
mistral-7b ERROR - check error output
traceback
2026-07-24 16:56:41.036762 [WARN] [/data/src/onnx/onnx_parser.cpp:296] Model has unbound symbolic dimension(s): batch_size, sequence_length. These default to 1 and may cause unexpected behavior. Try setting --dim-param @<name> <value> or --input-dim @<input> <dims> if program compilation fails.
2026-07-24 16:57:47.316941 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:57:47.317004 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution135262927453122.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
FLUX.1-schnell ERROR - check error output
traceback
2026-07-24 16:58:47.671176 [WARN] [/data/src/onnx/onnx_parser.cpp:296] Model has unbound symbolic dimension(s): B. These default to 1 and may cause unexpected behavior. Try setting --dim-param @<name> <value> or --input-dim @<input> <dims> if program compilation fails.
2026-07-24 16:58:51.255122 [ERROR] [/data/src/pass_manager.cpp:176] Error layout_convolution: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape
2026-07-24 16:58:51.255178 [ERROR] [/data/src/pass_manager.cpp:186] Dump: "/tmp/migraphx/layout_convolution135326865632842.mxr"
Traceback (most recent call last):
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in
main()
File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 246, in main
model.compile(
RuntimeError: /data/src/shape.cpp:1025: dyn_dims: SHAPE: dyn_dims() called on a static shape

@codecov

codecov Bot commented Jul 23, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

Additional details and impacted files
@@             Coverage Diff             @@
##           develop    #5086      +/-   ##
===========================================
- Coverage    92.90%   92.90%   -0.00%     
===========================================
  Files          603      603              
  Lines        32526    32528       +2     
===========================================
+ Hits         30217    30218       +1     
- Misses        2309     2310       +1     

see 2 files with indirect coverage changes

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 2 comments.

Comments suppressed due to low confidence (1)

CMakeLists.txt:475

  • The else-branch comment was changed to "Legacy ROCm packaging", but this is still the default/traditional packaging backend. Calling it "legacy" is potentially misleading in configuration output and future maintenance.
    # Legacy ROCm packaging.
    set(_MGX_PKG_NAME "MIGraphX")

Comment thread CMakeLists.txt Outdated
Comment on lines 451 to 454
set(MIGRAPHX_THEROCK_GPU_ARCH "" CACHE STRING
"TheRock GPU arch(es) for per-GPU package dependencies. \
Semicolon-separated list (e.g. gfx90a;gfx942;gfx950) for per-GPU deps, \
or empty for device-all meta-package deps.")
Comment thread CMakeLists.txt Outdated
Comment on lines +457 to +461
# Self-contained per-arch packages (amdrocm-migraphx-gfxNNN) and their per-GPU
# amdrocm-* deps are produced by the UFB Dockerfile cpack loop: one fat build,
# repackaged once per arch. Here we set only the base name + arch-neutral
# runtime dep; the device-all deps below are the standalone default and are
# overridden per-arch (amdrocm-dnn/blas-gfxNNN) at cpack time.
@kentqian kentqian changed the title [AIMIGRAPHX-1186] TheRock: deb/rpm per-arch meta packages for external ROCm deps (drop weak Recommends/Suggests) [AIMIGRAPHX-1186] TheRock packaging: amdrocm-migraphx naming + drop weak deps (per-arch packaging in UFB) Jul 23, 2026

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

Comment thread CMakeLists.txt
Comment on lines +453 to +457
# Only arch-neutral runtime deps are declared here (amdrocm-runtime), since
# that is the only component published as an arch-neutral package.
# The device libraries (MIOpen/rocBLAS) exist ONLY as per-arch packages
# (amdrocm-dnn-gfxNNN / amdrocm-blas-gfxNNN) with no arch-neutral variant,
# so they are intentionally omitted here. To produce a package with device
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

dependencies Pull requests that update a dependency file

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants