Use 64-bit warp mask for ROCm HIP compatibility

[harkgill-amd]

Same change that when into torch-scatter here rusty1s/pytorch_scatter@648dda3 from @Looong01.

ROCm 6.4.3+ __shfl_sync / __shfl_up_sync / __shfl_down_sync function templates now require the mask parameter to be a 64-bit integer (uint64_t).

Before this patch, we'd see static assertion failed: sizeof(unsigned int) == 8 when building for ROCm.

With this patch, build is successful.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 73.09%. Comparing base (2d55981) to head (593007c).
⚠️ Report is 15 commits behind head on master.
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #413   +/-   ##
=======================================
  Coverage   73.09%   73.09%           
=======================================
  Files          29       29           
  Lines        1171     1171           
=======================================
  Hits          856      856           
  Misses        315      315           

☔ 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.