RepoJournal
PyTorch @pytorch
90 wires and counting

$ follow PyTorch

Keep up with PyTorch in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-06-14
stories 36
GitHub RSS Archive Reviews Topics

© 2026 RepoJournal Home Showcase Explore How it works Privacy

$ the-wire · showcase

INDUCTOR DOUBLES DOWN ON PROFILING; C10D BACKENDS GET UNIFIED API

By RepoJournal · Filed · About PyTorch · Composed from the cited sources · methodology

tianyu-l yichuan-w xmfan
3 people shipped this
tianyu-l @tianyu-l 1 cited yichuan-w @yichuan-w 1 cited xmfan @xmfan 1 cited

PyTorch's compilation stack shipped major profiler improvements and collective communication overhaul overnight, with changes spanning kernel provenance tracking to 390K-entry cache optimization.

The Inductor profiler now attaches kernel provenance metadata to Chrome traces [1], threading Triton and extern-kernel context through debug paths so you can see exactly which source lines generated which CUDA kernels in your timeline. That lands alongside a sweeping c10d backend refactor [2] that introduces the torchcomms `_single` collective names on the C++ API while keeping all legacy names fully backward compatible, unifying how distributed training backends declare their capabilities. On the compiler front, Inductor's NVGEMM handler fixed a brutal performance bug [3] where the kernel cache search iterated through 390K candidates twice per GEMM operation; a new single-pass partition function cuts that overhead entirely. The same area shed dead code [4] where duplicate nvgemm_max_profiling configs were shadowing the canonical definition. In gradient optimization, linear_cross_entropy now stops materializing zero-filled gradients [5] for unused chunked-op outputs, cutting unnecessary backward overhead. TorchTitan's MoE layer got leaner: production models all shipped with `score_before_experts=False`, so [6] removed the dead `True` branch entirely. The repo also added a deterministic loss test [7] for Qwen3 with TP+EP configs to catch silent numerical corruption, and landed MinimalAsyncEP [8], a cudagraphable expert dispatcher with minimal kernel overhead and a ping-pong buffer pattern that avoids symm-mem copies during recompute. FBGEMM finished migrating TBE backward templates [9] to the new threshold-guard API and replaced deprecated PyTorch calls [10] across the fbgemm_gpu surface.

Action items

References

  1. [1] Add Inductor profiler timeline provenance (#186230) pytorch/pytorch ↗
  2. [2] Add _single c10d::Backend methods and migrate backends to them (#187140) (#187140) pytorch/pytorch ↗
  3. [3] [Inductor][NVGEMM] Avoid double iteration over kernel cache during choice enumeration (#185966) pytorch/pytorch ↗
  4. [4] [Inductor][NVGEMM] Remove duplicate nvgemm_max_profiling config (#185965) pytorch/pytorch ↗
  5. [5] linear_cross_entropy: do not materialize gradients for unused chunked-op outputs (#187219) pytorch/pytorch ↗
  6. [6] [MoE] Remove unused score_before_experts dispatcher flag ↗ pytorch/torchtitan
  7. [7] [MoE] Add TP+EP config to Qwen3 MoE deterministic loss test (FSDP4+TP2+EP4) ↗ pytorch/torchtitan
  8. [8] Add MinimalAsyncEP ↗ pytorch/torchtitan
  9. [9] Migrate TBE backward template to cap_grid_dim_x; clean up legacy get_max_thread_blocks helpers (#5853) ↗ pytorch/FBGEMM
  10. [10] Replace deprecated PyTorch APIs in fbgemm_gpu (#5896) ↗ pytorch/FBGEMM

$ ls pytorch/reviews/ # the step-back reads

the week in review Helion autotunes Metal, ending default 16x16x16 tiles the month in review ExecuTorch MoE quantization, Helion runtime splits, and a broader cleanup wave
Topics: Python AI / ML Full archive → Weekly reviews →

Quick answers

What shipped in PyTorch on June 14, 2026?
PyTorch's compilation stack shipped major profiler improvements and collective communication overhaul overnight, with changes spanning kernel provenance tracking to 390K-entry cache optimization. In total, 29 commits and 7 pull requests landed.
Who contributed to PyTorch on June 14, 2026?
3 developers shipped this update, including tianyu-l, yichuan-w, and xmfan.
What were the notable PyTorch updates?
Add Inductor profiler timeline provenance (#186230), Add _single c10d::Backend methods and migrate backends to them (#187140) (#187140), and [Inductor][NVGEMM] Avoid double iteration over kernel cache during choice enumeration (#185966).
♥

$ log @pytorch --recent

torchtitan fixes multi-node HF export, torch._native opens to non-aten namespaces 57 shipped Fake-tensor hardening, ROCm graph fixes, and an XNNPACK use-after-free 180 shipped CPU caching allocator teardown race fixed, torchtitan shares one training engine 67 shipped ROCm test timeouts traced to duplicated runs, and distributed/test_store drops 15 minutes 100 shipped

$ log --all # elsewhere on the wire

@openclaw Cold-start Gateway diagnostics stop crying wolf, doctor repairs plugin drift Agentic Coding Cline desktop bundles shrink, goose unblocks its own clicks

$ ls showcases/

Rails Django Laravel Node.js Go Vercel Vue.js Kubernetes HashiCorp Arch Linux Linux Anthropic

$ init your-desk

RepoJournal can write one of these for any repo you care about. Sign in with GitHub, pick your repos and sources, and your desk gets a wire like this one every morning.