io-aware

Tag

Cards List
#io-aware

@_avichawla: Researchers made KMeans 200x faster. And the new technique also beats approaches like cuML and FAISS. Flash-KMeans is a…

X AI KOLs Timeline · 2026-06-16 Cached

Flash-KMeans is an IO-aware implementation of exact KMeans that redesigns the algorithm around modern GPU bottlenecks, achieving 33x speedup over cuML and 200x over FAISS by eliminating redundant memory reads and writes.

0 favorites 0 likes
#io-aware

TIDE: Efficient and Lossless MoE Diffusion LLM Inference with I/O-aware Expert Offload

Hugging Face Daily Papers · 2026-05-19 Cached

TIDE is a lossless inference system for diffusion large language models that leverages temporal stability of expert activations to reduce I/O overhead and computation, achieving up to 1.4-1.5x throughput improvements on single GPU-CPU systems.

0 favorites 0 likes
← Back to home

Submit Feedback