feature: add support for SMDDP collectives to smdataparallel runner

[vishwakaria]

Description of changes: Add support for SMDDP collectives in PT DDP distribution via a new parameter in the dictionary:

communication_options: {
                        "backend": "nccl",  #default value is auto
                       }

The distribution will set a configuration parameter called sagemaker_communication_backend. If the value is auto, we will preload libsmddp which has the Sagemaker optimized implementation of AllReduce. If the value is nccl, we will just use the nccl-allReduce implementation.

Testing done:

• Added unit tests

• Created a custom docker image and ran successful training

• Related PySDK PR: https://github.com/aws/sagemaker-python-sdk/pull/3490

Merge Checklist

Put an x in the boxes that apply. You can also fill these out after creating the PR. If you're unsure about any of them, don't hesitate to ask. We're here to help! This is simply a reminder of what we are going to look for before merging your pull request.

General

• [X] I have read the CONTRIBUTING doc
• [X] I used the commit message format described in CONTRIBUTING
• [ ] I have used the regional endpoint when creating S3 and/or STS clients (if appropriate)
• [X] I have updated any necessary documentation, including READMEs

Tests

• [X] I have added tests that prove my fix is effective or that my feature works (if appropriate)
• [X] I have checked that my tests are not configured for a specific region or account (if appropriate)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-training-toolkit-pr
• Commit ID: a3f05c9e004440589f0fdaaffec6af9856fc6b63
• Result: FAILED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-training-toolkit-pr
• Commit ID: 92c983db521e0f2d409bda2d1bd6544751bf37f9
• Result: FAILED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-training-toolkit-pr
• Commit ID: 61b3a32e8a6bb8636e35a71a290438191e4121c6
• Result: FAILED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-training-toolkit-pr
• Commit ID: 5b0826cd9c80fc157173f2a4dd7be744bab03fa2
• Result: SUCCEEDED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[satishpasumarthi]

Please follow the commit message style as per https://github.com/aws/sagemaker-training-toolkit/blob/master/CONTRIBUTING.md#committing-your-change If you are done with your changes, see if you can squash them into one @vishwakaria

[vishwakaria]

Please follow the commit message style as per https://github.com/aws/sagemaker-training-toolkit/blob/master/CONTRIBUTING.md#committing-your-change If you are done with your changes, see if you can squash them into one @vishwakaria

Addressed your comments and squashed all commits. Can you take another look @satishpasumarthi? Thank you.

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-training-toolkit-pr
• Commit ID: 127b0b133e600eacf4af811c91c4c5f0a9de7f6a
• Result: SUCCEEDED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[sagemaker-bot]

AWS CodeBuild CI Report

• CodeBuild project: sagemaker-training-toolkit-pr
• Commit ID: 5e9e42caa9d35766e56a009f3da08b3355dada89
• Result: SUCCEEDED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository