abft

Here are 2 public repositories matching this topic...

GautamTalksDev / assay-gpu

Empirically calibrated false-positive bound for one-sided checksum ABFT on bfloat16 GEMM, where the classical τ = Ku threshold is undefined. Catches 76% of exponent-bit flips against a pre-registered 90% bar.

fault-tolerance reproducible-research gpu reliability cuda pytorch gemm error-detection abft pre-registration bfloat16 numerical-computing extreme-value-theory silent-data-corruption

Updated Oct 6, 2026
Python

ArchanaChetan07 / Fault-Tolerant-Distributed-AI-Training-System

Star

Fault-tolerant distributed training: ABFT-protected GEMMs, elastic checkpoint recovery, and PyTorch DDP under node failure and SEU bit-flips

distributed-systems machine-learning deep-learning reliability-engineering fault-tolerance pytorch high-performance-computing mlops abft checkpointing

Updated Aug 5, 2026
Python

Add this topic to your repo

To associate your repository with the abft topic, visit your repo's landing page and select "manage topics."

Learn more