Fix schedule_all chunking for small ranges

[fallintoplace]

Summary

Fix static_thread_pool::schedule_all batching so small non-empty ranges always make progress when the pool has more workers than items. The batching code now asserts the existing non-zero worker invariant, clamps non-empty chunk sizes to at least one item, and resets the pending task count after each intermediate flush.

Testing

• cmake --build build --target test.exec -j 8
• ./build/test/exec/test.exec "schedule_all on static_thread_pool handles ranges smaller than available parallelism"
• ./build/test/exec/test.exec "[types][static_thread_pool]"
• ./build/test/exec/test.exec

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.