[develop-upstream-QA-rocm61] terminate called after throwing an instance of 'std::system_error' workaround

ROCm / tensorflow-upstream

TensorFlow ROCm port

https://tensorflow.org

Apache License 2.0

684 stars 93 forks source link

[develop-upstream-QA-rocm61] terminate called after throwing an instance of 'std::system_error' workaround #2478

Closed pemeliya closed 6 months ago

pemeliya commented 6 months ago

This PR fixed sleeping threads problem: https://github.com/tensorflow/tensorflow/issues/62466 already merged to upstream: https://github.com/ROCm/tensorflow-upstream/pull/2370

The respective Jira issue is: https://ontrack-internal.amd.com/browse/SWDEV-448642

i-chaochen commented 6 months ago

LGTM. Do we need to get this fix to r2.13, r2.14, r2.15?

echo this @pemeliya I think we need this to r2.13, r2.14 and r2.15 rocm-enhanced as this file is same across all versions, the problem would need this patch.