PyTorch 2.15 adds NCCL backend to mitigate single‑GPU failures

The 2.15 release of PyTorch rewrites the NCCL backend to improve fault diagnosis and recovery, aiming to prevent a single bad GPU from forcing a full training restart.

Why it matters

Background only: no immediate product decision indicated.

For product managers

For your product

What does this mean for your product?

Try it on your product

Read at The Register

More from this issueAll AI news from 7 Nov →

Thinking Machines Lab releases Inkling 2 Small 594B MoE with 17B activeAvailable now

OpenAI valued at $1.25 trillion after $11 billion funding round

Amazon launches text‑based Alexa shopping assistant for US customersAnnounced

Ricursive Intelligence releases Ripple-65B Preview, tops open US models on TerminalBenchAvailable now

Get this every morning, written for your product.

Deltaday reads the whole AI news cycle and tells you what changed, whether it's real yet, and what to do about it.

Start free See how it works