Repository navigation
test(miopen): exclude failing LayerNormBwd FP16 cases on gfx110X and gfx1151 (#15056) - #15085
Closed
aalabsi-amd wants to merge 1 commit into
Closed
aalabsi-amd wants to merge 1 commit into
aalabsi-amd wants to merge 1 commit into
Conversation
…gfx1151 (#15056) Full/GPU_LayerNormBwdTest_FP16.LayerNormTestBw/124 (mode 0) and /127 (mode 1), N:1 C:2048 H:4 W:32 dim:3 layout:1, fail deterministically with "Error dw beyond tolerance Error:0.0116, Threshold x 2: 0.00195" (layernorm.hpp:524) on Linux and Windows gfx110X and gfx1151 since #13107 added */GPU_LayerNormBwdTest_* to the comprehensive tier. Every attempt of develop nightlies 41764975629, 41793654564, 41764975562, 41793655004 failed. gfx120X passes. Exclude the two exact cases in exclude_gpu_gfx110X and exclude_gpu_gfx115X (not gfx11_common_patterns, which gfx120X shares). Jenkins runs no gfx110X/gfx115X tests, so no add_gtest_negative_filter is needed. Validated with yaml_select.py diff vs origin/develop: comprehensive and full drop exactly these 2 cases on gfx1101/gfx1151 Linux and Windows; no change on gfx1201, gfx942 or the base suite. Remove when #15056 closes.
aalabsi-amd
added
project: miopen
ci:gfx1151
✅ 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. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Exclude the two exact cases LayerNormTestBw/124 and LayerNormTestBw/127 from nightly runs in exclude_gpu_gfx110X and exclude_gpu_gfx115X (not gfx11_common_patterns, which gfx120X shares). Jenkins runs no gfx110X/gfx115X tests, so no add_gtest_negative_filter is needed.
Technical Details
Full/GPU_LayerNormBwdTest_FP16.LayerNormTestBw/124 (mode 0) and /127 (mode 1), N:1 C:2048 H:4 W:32 dim:3 layout:1, fail deterministically with "Error dw beyond tolerance Error:0.0116, Threshold x 2: 0.00195" (layernorm.hpp:524) on Linux and Windows gfx110X and gfx1151 since #13107 added /GPU_LayerNormBwdTest_ to the comprehensive tier. Every attempt of develop nightlies 41764975629, 41793654564, 41764975562, 41793655004 failed. gfx120X passes.
Test Plan
Validated with yaml_select.py diff vs origin/develop: comprehensive and full drop exactly these 2 cases on gfx1101/gfx1151 Linux and Windows; no change on gfx1201, gfx942 or the base suite. Remove when #15056 closes.
Test Result
Submission Checklist