Skip to content
This repository has been archived by the owner on May 17, 2022. It is now read-only.

Implemented Torch-UCC CI (WIP) #33

Open
wants to merge 9 commits into
base: ucc
Choose a base branch
from
Open

Implemented Torch-UCC CI (WIP) #33

wants to merge 9 commits into from

Conversation

artemry-nv
Copy link

@artemry-nv artemry-nv commented Feb 17, 2021 •
edited
Loading

Changes:

  • Prepared base docker environment
  • Added building scenarios

@artemry-nv artemry-nv changed the title Implemented Torch-UCC CI Implemented Torch-UCC CI (WIP) Feb 17, 2021
@artemry-nv
Copy link
Author

@mike-dubman @Sergei-Lebedev please grant me permissions to add reviewers.

@artemry-nv
Removed some debug stuff
099ca8b 
Signed-off-by: artemry-nv <[email protected]>
mike-dubman
mike-dubman reviewed Feb 18, 2021
View reviewed changes
mike-dubman
mike-dubman reviewed Feb 18, 2021
View reviewed changes
mike-dubman
mike-dubman reviewed Feb 18, 2021
View reviewed changes
@mike-dubman
Copy link
Contributor

@mike-dubman @Sergei-Lebedev please grant me permissions to add reviewers.

done. plz check your inbox and accept invite

@artemry-nv
Copy link
Author

@mike-dubman @Sergei-Lebedev please grant me permissions to add reviewers.

done. plz check your inbox and accept invite

Don't see any invitations - could you please re-send it?

@artemry-nv
Reworked for CentOS 8
e5553df 
Signed-off-by: artemry-nv <[email protected]>
@artemry-nv artemry-nv force-pushed the artemry-nv/torch-ucc-ci branch from 32591b2 to e5553df Compare March 15, 2021 06:40
lennybe
lennybe reviewed Mar 15, 2021
View reviewed changes
@artemry-nv artemry-nv changed the base branch from master to ucc March 29, 2021 19:32
mike-dubman
mike-dubman reviewed Mar 30, 2021
View reviewed changes
.ci/scripts/run_tests_torch_ucc.sh
TORCH_UCC_SRC_DIR_WITH_XCCL_SUPPORT="${TORCH_UCC_SRC_DIR}_xccl"

echo "UCC barrier"
/bin/bash ${TORCH_UCC_SRC_DIR_WITH_XCCL_SUPPORT}/test/start_test.sh ${TORCH_UCC_SRC_DIR_WITH_XCCL_SUPPORT}/test/torch_barrier_test.py --backend=gloo
Copy link
Contributor

Choose a reason for hiding this comment

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

why backend=gloo?

Copy link
Author

Choose a reason for hiding this comment

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

@Sergei-Lebedev could you please comment?

Copy link
Collaborator

Choose a reason for hiding this comment

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

backend option sets backend to compare with, i.e. ucc collective will be compared with gloo collective for correctness check

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@Sergei-Lebedev Sergei-Lebedev Sergei-Lebedev left review comments

@lennybe lennybe lennybe left review comments

@mike-dubman mike-dubman mike-dubman approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@artemry-nv @mike-dubman @Sergei-Lebedev @lennybe