Arm backend: Streamline MLSDK deps and standardise setup logging

[ArmRyan]

• ensure download_ai_mlsdk_manifest only refreshes when URL/tag change (avoids needless MLSDK downloads) and uses consistent cleanup
• centralise log_step usage, trim noisy logs, and add get_parallel_jobs so repo sync/build steps pick the right concurrency everywhere
• centralise log_step and parallelism across other setup scripts also.

Change-Id: I4edad30c7ced94afda0a217e3f7bf291e9e5e2f8

cc @freddan80 @per @zingo @oscarandersson8218 @digantdesai

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15189

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit 2a66367 with merge base 1f114f1 ():

NEW FAILURES - The following jobs have failed:

• pull / test-samsung-models-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 8cb103d68613107c7504cce1f3b724d8cc8bc8691b1b1231b7c14c3af622d20e /exec failed with exit code 1
• Test CUDA Builds / export-voxtral-cuda-artifact / linux-job (gh)
  RuntimeError: Command docker exec -t 7fb3a281d7baaffc5239e21247ca3232d5e9f998e42845a91b7a00fef2896caf /exec failed with exit code 2

This comment was automatically generated by Dr. CI and updates every 15 minutes.