Issues

• Separating out the ID-Model from the rest of the nvFuser code base.

  NVIDIA/Fuser#6029 · Filter by author kevinstephano opened on Mar 6, 2026

• [Multidevice] Fused comm+compute kernel generation for multi-GPU operations

  NVIDIA/Fuser#6003 · Filter by author samnordmann opened on Feb 23, 2026
  4

• Resharding set with alternate DIDx and DIDy

  Multi-GPU
  Bug

  NVIDIA/Fuser#6001 · Filter by author wujingyue opened on Feb 23, 2026
  1

• Reduce memory usage in distributed triangle updates

  Multi-GPU
  NVIDIA/Fuser#5942 · Filter by author wujingyue opened on Feb 8, 2026

• Improve sharding propagation for triangle updates outgoing

  Multi-GPU
  Enhancement

  NVIDIA/Fuser#5901 · Filter by author wujingyue opened on Jan 30, 2026
  8

• Should use pointwise scheudler but transpose scheduler is used

  NVIDIA/Fuser#5883 · Filter by author liqiangxl opened on Jan 27, 2026
  1

• unique_ptr axioms in Ir Containers

  NVIDIA/Fuser#5868 · Filter by author mdavis36 opened on Jan 23, 2026
  

• Optimize non-circular buffered TMA loads

  NVIDIA/Fuser#5858 · Filter by author liqiangxl opened on Jan 21, 2026
  1

• Python/CMake/Env build option refresh

  NVIDIA/Fuser#5824 · Filter by author mdavis36 opened on Jan 14, 2026
  

• BFS traversal failure in IndexLowering pass

  NVIDIA/Fuser#5779 · Filter by author liqiangxl opened on Jan 8, 2026
  1

• Keep track of extent tensor dependency with RaggedIterDomain

  tensor list
  NVIDIA/Fuser#5721 · Filter by author naoyam opened on Dec 19, 2025
  1

  

• Inconsistent values for the same extent Val*

  NVIDIA/Fuser#5710 · Filter by author zasdfgbnm opened on Dec 18, 2025
  5

• Sub-optimal communication in transformer TP + SP backward block

  NVIDIA/Fuser#5673 · Filter by author Priya2698 opened on Dec 12, 2025

• Potential correctness issue with stmatrix

  Bug

  NVIDIA/Fuser#5657 · Filter by author naoyam opened on Dec 10, 2025

• Reconsider CompiledKernel compilation and ownership.

  NVIDIA/Fuser#5652 · Filter by author mdavis36 opened on Dec 9, 2025
  

• Debug output for Host IR JIT LaunchKernel

  NVIDIA/Fuser#5651 · Filter by author mdavis36 opened on Dec 9, 2025
  

• hir::LaunchKernel recompilation

  NVIDIA/Fuser#5625 · Filter by author mdavis36 opened on Dec 4, 2025
  

• hir::LaunchKernel Intermediate Buffer Support

  NVIDIA/Fuser#5624 · Filter by author mdavis36 opened on Dec 4, 2025
  

• codegen issue with allocation domain on cache

  allocation domain
  NVIDIA/Fuser#5611 · Filter by author jjsjann123 opened on Dec 2, 2025
  5

• RFC: Rename FusionDefinition.execute/manual_execute for clarity and add documentation on usage

  developer efficiencyDirect BindingsPython API
  Enhancement

  NVIDIA/Fuser#5582 · Filter by author IvanYashchuk opened on Nov 23, 2025
  1

  

• Unnecessary local mempcy in postAllReduce

  Multi-GPU
  NVIDIA/Fuser#5567 · Filter by author wujingyue opened on Nov 20, 2025

• Segmentation error in test_cross_entropy_bwd_benchmark[executor='thunder'-variation='hf_phi3']

  Python Benchmarks
  NVIDIA/Fuser#5542 · Filter by author rdspring1 opened on Nov 17, 2025
  2

• test_many_pointwise_ops_host.py is missing from cpu host latency in nightly dashboard.

  Python Benchmarks
  NVIDIA/Fuser#5531 · Filter by author rdspring1 opened on Nov 14, 2025
  2

  

• MultiDeviceExecutor ignores stream-parallel allocation IterDomains.

  NVIDIA/Fuser#5525 · Filter by author wujingyue opened on Nov 14, 2025
  3

• layout op does not work with tensor metadata

  NVIDIA/Fuser#5500 · Filter by author zasdfgbnm opened on Nov 11, 2025