-
Notifications
You must be signed in to change notification settings - Fork 27.4k
[aarch64 linux ci] missing sccache binaries for aarch64 linux platform #121559
Copy link
Copy link
Closed
Labels
module: armRelated to ARM architectures builds of PyTorch. Includes Apple M1Related to ARM architectures builds of PyTorch. Includes Apple M1triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module
Description
🐛 Describe the bug
Currently pytorch CI build infrastructure maintains sccache binaries only for x86 architecture. We need the aarch64 linux binaries to speed up the aarch64 ci workflows. Currently aarch64 linux workflow is crashing if sccache is enabled because x86 binaries are pulled.
sccache_epilogue' EXIT
++ [[ -n 1 ]]
++ echo 'Skipping sccache server initialization, setting environment variables'
Skipping sccache server initialization, setting environment variables
++ export SCCACHE_IDLE_TIMEOUT=0
++ SCCACHE_IDLE_TIMEOUT=0
++ export SCCACHE_ERROR_LOG=/var/lib/jenkins/sccache_error.log
++ SCCACHE_ERROR_LOG=/var/lib/jenkins/sccache_error.log
++ export RUST_LOG=sccache::server=error
++ RUST_LOG=sccache::server=error
++ sccache --zero-stats
.ci/pytorch/common-build.sh: line 47: /opt/cache/bin/sccache: cannot execute binary file: Exec format error
Versions
aarch64 linux runner:
ami-id: ami-0b1c2e7eafac77408
instance-id: i-0fbadc66663ad3bdc
instance-type: t4g.2xlarge
system info Linux ip-10-0-42-211.ec2.internal 4.14.336-[25](https://github.com/pytorch/pytorch/actions/runs/8164152112/job/22318870156#step:4:27)6.559.amzn2.aarch64 #1 SMP Tue Feb 13 09:50:21 UTC 2024 aarch64 aarch64 aarch64 GNU/Linux
cc @malfet
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
module: armRelated to ARM architectures builds of PyTorch. Includes Apple M1Related to ARM architectures builds of PyTorch. Includes Apple M1triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module