Search results

[ContextParallel][benchmark] Add sparsity benchmark over different load-balance policies

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Status: Draft (not ready).

pytorch/pytorchnumber 163806

#163806 In pytorch/pytorch;

· XilunWu opened on Sep 24, 2025

[ContextParallel] add process-time based Round-Robin load-balance to CP

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Status: Draft (not ready).

pytorch/pytorchnumber 163617

#163617 In pytorch/pytorch;

· XilunWu opened on Sep 23, 2025

[ContextParallel][example] FLOPS measure

ciflow/inductor

module: context parallelPyTorch Context ParallelPyTorch Context Parallel

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

topic: not user facingtopic categorytopic category

Status: Draft (not ready).

pytorch/pytorchnumber 163053

#163053 In pytorch/pytorch;

· XilunWu opened on Sep 16, 2025

[WIP] change DCP code to use the new util

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

release notes: distributed (checkpoint)

Status: Draft (not ready).

pytorch/pytorchnumber 161775

#161775 In pytorch/pytorch;

· XilunWu opened on Aug 29, 2025

[do NOT land] replace _StridedShard w/ sharding order in FSDP2

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

release notes: distributed (fsdp)release notes categoryrelease notes category

Status: Draft (not ready).

pytorch/pytorchnumber 161770

#161770 In pytorch/pytorch;

· XilunWu opened on Aug 29, 2025

[WIP] add util for computing local shape and global offset from DTensorSpec with extra tensor_sharding field

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Status: Draft (not ready).

pytorch/pytorchnumber 161704

#161704 In pytorch/pytorch;

· XilunWu opened on Aug 28, 2025

[ContextParallel] add load-balance interface to context_parallel()

ciflow/inductor

module: context parallelPyTorch Context ParallelPyTorch Context Parallel

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

topic: not user facingtopic categorytopic category

Status: Open (in progress).

pytorch/pytorchnumber 161062

#161062 In pytorch/pytorch;

· XilunWu opened on Aug 20, 2025

[Do Not Land][CP] sdpa debugging

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Status: Draft (not ready).

pytorch/pytorchnumber 160561

#160561 In pytorch/pytorch;

· XilunWu opened on Aug 13, 2025

[DTensor] migrate prop_index (sharding prop rule for op aten.index.Tensor) to sharding prop strategy

ciflow/inductor

module: dtensordistributed tensor tagdistributed tensor tag

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

topic: not user facingtopic categorytopic category

Status: Draft (not ready).

pytorch/pytorchnumber 159003

#159003 In pytorch/pytorch;

· XilunWu opened on Jul 24, 2025

[Tracker] AutoParallel's feature request to DTensor

module: dtensordistributed tensor tagdistributed tensor tag

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Status: Open.

#156217 In pytorch/pytorch;

· XilunWu opened on Jun 17, 2025

draft: [cp] context_parallel + flex_attention_backward using torch_function and autograd function

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Stale

Status: Draft (not ready).

pytorch/pytorchnumber 155970

#155970 In pytorch/pytorch;

· XilunWu opened on Jun 13, 2025

draft: [cp] context_parallel + flex_attention using torch_function and autograd function

ciflow/inductor

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

Stale

Status: Draft (not ready).

pytorch/pytorchnumber 155962

#155962 In pytorch/pytorch;

· XilunWu opened on Jun 13, 2025