feat(subsystembenchmarks): add Ray Data checkpoint load benchmark

[Yonghui-Lee]

Depends on #1043

This PR adds the checkpoint load subsystem benchmark for Ray Data and Ray Train PyTorch workloads.

It benchmarks loading single-node and distributed checkpoints across FSDP, DDP, and Tensor Parallel + Data Parallel strategies, including cross-topology resharding.

Key Changes

Ray Checkpoint Load Driver:

• Implemented RayCheckpointLoadWorker actor executing distributed checkpoint load operations via PyTorch Distributed Checkpoint (DCP) and ray.train.Checkpoint.
• Implemented RayCheckpointSetupWorker actor that pre-populates the source checkpoint on GCS using the case's setup topology (setup_world_size, setup_tensor_parallel_size, setup_data_parallel_size).
• Implemented RayCheckpointReadDriver managing cluster initialization, fixture setup, and benchmark execution.

[gemini-code-assist]

Code Review

This pull request introduces a new subsystem benchmark suite for Ray checkpointing (ray_data), which includes drivers, configurations, and tests for both checkpoint read and write scenarios. It also updates the Cloud Build configuration to support a configurable boot disk size. The review feedback suggests two improvements: logging warnings instead of silently suppressing exceptions when verifying staging directories in common.py, and simplifying the redundant parameter resolution logic in read/driver.py by directly accessing attributes already handled by CheckpointParameters.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.18%. Comparing base (758f35c) to head (225597f).
⚠️ Report is 4 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1047      +/-   ##
==========================================
+ Coverage   90.10%   90.18%   +0.08%     
==========================================
  Files          16       16              
  Lines        3687     3740      +53     
==========================================
+ Hits         3322     3373      +51     
- Misses        365      367       +2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.