PyTorch Issue: Inconsistent Randomness Across Devices with Deterministic Algorithms
rice_fry · x · 2026-08-25
A PyTorch GitHub Issue (#84234) highlights that random number generation remains inconsistent across different GPUs even when usedeterministicalgorithms is enabled.
The issue argues that as a research platform, PyTorch should support reproducible results across different devices (CPU or physical GPU) given the same seed, PyTorch version, and CUDA version. The current inconsistency hinders experiment reproducibility.
More from coding & agent
- Engineer: my job went from silently thinking at a computer to constantly talking to it — KevinNaughtonJr · 2026-10-16
- Vercel Goes Full Backend: Python GA, Containers, and Daemons — cramforce · 2026-10-16
- WorldMemArena Accepted at NeurIPS: Memory Storage ≠ Memory Use — xwang_lk · 2026-10-16
- Vercel Ship Demos Container Deploys for Go, Ruby, Java and More — edwin · 2026-10-16
- TRL adds day-0 support for fine-tuning Muse Spark Open's 816B MoE with LoRA on 64 H100s — QGallouedec · 2026-10-16
- Databricks engineers land impactful PRs on day 2 with AI dev tools, vs week 3 before — Yuchenj_UW · 2026-10-16