Added support for topographical ordering of hostnames in mpi run

[harishvs]

Issue #, if available:

Description of changes:
Added support for topographical ordering of hostnames in mpi run
and made the slurm sbatch scripts more generic and added convenience scripts to convert nccl output to excel

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[harishvs]

@amanshanbhag @nghtm

[KeitaW]

Thanks for the PR. The tool looks useful, but shouldn't we create a separate directory for it instead of overwriting the existing Slurm example? Few workshops depend on the test case (e.g., https://catalog.workshops.aws/ml-on-aws-parallelcluster/en-US, https://catalog.workshops.aws/sagemaker-hyperpod/en-US). Also, we would like to show how to run the equivalent tests on both Slurm/Kubernetes using the current examples.
Left some initial comments.

[harishvs]

@KeitaW thanks for the detailed review. I moved my files to a v2 directory under slurm to preserve the the previous versions. I have addressed all your comments. please take a look.

[nghtm]

confirming @KeitaW and @amanshanbhag assigned for review

[KeitaW]

Thanks for updating them. Few more comments. Could you please name the subdirectory in a bit more descriptive name instead of v2 (thanks for moving the scripts into the directory)? Maybe topology-aware-nccl-tests?

[KeitaW]

Suggested change

	docker build -f nccl-tests.Dockerfile \
	--build-arg="EFA_INSTALLER_VERSION=${EFA_INSTALLER_VERSION}" \
	--build-arg="AWS_OFI_NCCL_VERSION=${AWS_OFI_NCCL_VERSION}" \
	--build-arg="NCCL_VERSION=${NCCL_VERSION}" \
	--build-arg="NCCL_TESTS_VERSION=${NCCL_TESTS_VERSION}" \
	-t ${CONTAINER_IMAGE_NAME_TAG} \
	.
	cd ../..
	docker build -f nccl-tests.Dockerfile \
	--build-arg="EFA_INSTALLER_VERSION=${EFA_INSTALLER_VERSION}" \
	--build-arg="AWS_OFI_NCCL_VERSION=${AWS_OFI_NCCL_VERSION}" \
	--build-arg="NCCL_VERSION=${NCCL_VERSION}" \
	--build-arg="NCCL_TESTS_VERSION=${NCCL_TESTS_VERSION}" \
	-t ${CONTAINER_IMAGE_NAME_TAG} \
	.
	cd -

[harishvs]

In the terminal, we have not asked the user to navigate to awsome-distributed-training/micro-benchmarks/nccl-tests/slurm/topology-aware-nccl-tests prior to this step, so we assume that the user is at the project root dir in the terminal and direct accordingly.

so i added
#Navigate to the slurm directory:
cd micro-benchmarks/nccl-tests/slurm/

docker build -f nccl-tests.Dockerfile
--build-arg="EFA_INSTALLER_VERSION=${EFA_INSTALLER_VERSION}"
--build-arg="AWS_OFI_NCCL_VERSION=${AWS_OFI_NCCL_VERSION}"
--build-arg="NCCL_VERSION=${NCCL_VERSION}"
--build-arg="NCCL_TESTS_VERSION=${NCCL_TESTS_VERSION}"
-t ${CONTAINER_IMAGE_NAME_TAG}
.

I also updated line 106 accordingly. 

Navigate to the topology-aware-nccl-tests directory:
```bash
cd topology-aware-nccl-tests