Set lower CPU usage for test_shared_process_pool.py to avoid slowing down the test

[yczhang-nv]

Description

test_shared_process_pool.py is taking longer than expected when running on remote-ci, causing multiple ci pipelines failed due to (1h) timeout.
This is probably because:

• The number of workers of SharedProcessPool is set to 0.5 * os.sched_getaffinity(0) by default. If the test is running on a machine that has 64 cores, it will maintain 32 sub-processes.
• For each unit test, the process pool need to create all the sub-processes and then join() all of them when tear down, which might be slow when the number of processes is large.
• Running the test in remote-ci is slower than running locally, probably because there are multiple ci jobs running on the same machine

The test should run faster by setting the number of workers to 0.1 * os.sched_getaffinity(0) and it should avoid ci from being blocked by timeout.

Closes #1936

By Submitting this PR I confirm:

• I am familiar with the Contributing Guidelines.
• When the PR is ready for review, new or existing tests cover these changes.
• When the PR is ready for review, the documentation is up to date with these changes.

[yczhang-nv]

/merge