--- title: PyTorch distributed is changing and TorchComms is why description: TorchComms is PyTorch's new backend layer designed to replace the aging c10d backend substrate for distributed training and inference at scale. Rather than... canonical: https://daily.dev/posts/pytorch-distributed-is-changing-and-torchcomms-is-why-bbyvls17k twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: PyTorch distributed is changing and TorchComms is why | daily.dev og:description: TorchComms is PyTorch's new backend layer designed to replace the aging c10d backend substrate for distributed training and inference at scale. Rather than... og:url: https://daily.dev/posts/pytorch-distributed-is-changing-and-torchcomms-is-why-bbyvls17k og:image: https://api.daily.dev/og/posts/bByvlS17k.png og:image:alt: PyTorch distributed is changing and TorchComms is why og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # PyTorch distributed is changing and TorchComms is why **[Red Hat Developer](https://daily.dev/sources/rhdev)** · 6 min read · 0 upvotes · 0 comments ## Summary TorchComms is PyTorch's new backend layer designed to replace the aging c10d backend substrate for distributed training and inference at scale. Rather than replacing the user-facing torch.distributed API, it introduces explicit per-device communicators that plug back into existing ProcessGroup and DeviceMesh infrastructure via a wrapper path. Key benefits include dramatically faster process-group initialization (e.g., 265s down to 24s at 96,000 GPUs), significant inference decode-time improvements (up to 83% at 16 hosts), and support for communication primitives like windows and RMA that don't fit the old backend abstraction. TorchTitan already exposes a --comm.mode torchcomms flag, and migration requires minimal code changes. The ncclx backend path is not yet fully wired through init_process_group, but the core architecture is already visible in the open source tree. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Similar posts on daily.dev - [torchcomms: a modern PyTorch communications API – PyTorch](https://daily.dev/posts/torchcomms-a-modern-pytorch-communications-api-pytorch-ge8j0flky) · PyTorch · 0 upvotes · 0 comments - [PyTorch 2.14 Release Blog – PyTorch](https://daily.dev/posts/pytorch-2-14-release-blog-pytorch-civrznqmw) · PyTorch · 2 upvotes · 0 comments - [MPI-powered gradient synchronization in PyTorch distributed training](https://daily.dev/posts/mpi-powered-gradient-synchronization-in-pytorch-distributed-training-rfcnxqiyn) · Red Hat Developer · 0 upvotes · 0 comments --- Tags: [#machine-learning](https://daily.dev/tags/machine-learning), [#gpu](https://daily.dev/tags/gpu), [#distributed-systems](https://daily.dev/tags/distributed-systems), [#pytorch](https://daily.dev/tags/pytorch) [View this post on daily.dev](https://daily.dev/posts/pytorch-distributed-is-changing-and-torchcomms-is-why-bbyvls17k) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"PyTorch distributed is changing and TorchComms is why","url":"https://daily.dev/posts/pytorch-distributed-is-changing-and-torchcomms-is-why-bbyvls17k","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/pytorch-distributed-is-changing-and-torchcomms-is-why-bbyvls17k"},"datePublished":"2026-07-20T03:32:35.070Z","dateModified":"2026-07-20T03:32:59.510Z","description":"TorchComms is PyTorch's new backend layer designed to replace the aging c10d backend substrate for distributed training and inference at scale. Rather than...","image":"https://media.daily.dev/image/upload/s--58gMhC4P--/f_auto/v1722860399/public/Placeholder%2012","thumbnailUrl":"https://media.daily.dev/image/upload/s--58gMhC4P--/f_auto/v1722860399/public/Placeholder%2012","isAccessibleForFree":true,"articleSection":"Red Hat Developer","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Red Hat Developer","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/2a8895f3216c4900bdc455585a4ad3e4","url":"https://daily.dev/sources/rhdev"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/pytorch-distributed-is-changing-and-torchcomms-is-why-bbyvls17k","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"machine-learning,gpu,distributed-systems,pytorch","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Red Hat Developer","item":"https://daily.dev/sources/rhdev"},{"@type":"ListItem","position":3,"name":"PyTorch distributed is changing and TorchComms is why"}]} ```