Issues

• Add explanation of schedule torch profile configuration to readme

  LBANN/ScaFFold#90 · Filter by author michaelmckinsey1 opened on Jul 30, 2026
  

• Deprecate the "sweep" feature of benchmark.py

  LBANN/ScaFFold#89 · Filter by author PatrickRMiles opened on Jul 30, 2026
  

• Sharded volumegen

  LBANN/ScaFFold#64 · Filter by author PatrickRMiles opened on May 1, 2026
  

• Improve GPU utilization

  LBANN/ScaFFold#62 · Filter by author PatrickRMiles opened on Apr 30, 2026
  

• Pass fused arg to optimizer construction

  LBANN/ScaFFold#58 · Filter by author PatrickRMiles opened on Apr 23, 2026
  1

  

• Disable implicit gemm kernels to fix gridwiseop warning

  LBANN/ScaFFold#42 · Filter by author PatrickRMiles opened on Apr 2, 2026
  

• Test perf difference with data loading changes w/ reduced tensor copies

  LBANN/ScaFFold#38 · Filter by author PatrickRMiles opened on Mar 26, 2026

• Warn users if only naive kernel(s) is/are returned from search

  LBANN/ScaFFold#37 · Filter by author PatrickRMiles opened on Mar 26, 2026
  1

• Does dataloader num_workers actually impact data loading speed?

  LBANN/ScaFFold#36 · Filter by author PatrickRMiles opened on Mar 26, 2026

• Add Michael to authors list in pyproject.toml

  LBANN/ScaFFold#33 · Filter by author PatrickRMiles opened on Mar 25, 2026

• Slow performance on ROCm system

  LBANN/ScaFFold#18 · Filter by author rxy900 opened on Feb 20, 2026
  2

• Write stats summary to file when run completes

  enhancement
  LBANN/ScaFFold#13 · Filter by author PatrickRMiles opened on Jan 30, 2026
  

• Document min/max number of DistConv shards required at each scale

  documentation
  LBANN/ScaFFold#4 · Filter by author PatrickRMiles opened on Jan 22, 2026