paper-with-me

홈 › Papers

Flash-GMM: A Memory-Efficient Kernel for Scalable Soft Clustering

2026-06-09 · Gal Bloch, Ariel Gera, Matan Orbach, Ohad Eytan, Assaf Toledo arxiv

We present \textbf{Flash-GMM}, a fused Triton kernel for efficient computation of Gaussian Mixture Models (GMMs) over large-scale data in a single GPU pass. By eliminating the need to materialize the full responsibility matrix in GPU memory, Flash-GMM achieves a \textbf{20$\times$} speedup over existing implementations and enables training on datasets more than \textbf{100$\times$} larger than previously feasible on one device. To demonstrate its impact, we integrate Flash-GMM into the IVF coarse quantizer for approximate nearest-neighbor (ANN) search. We show that soft GMM clustering is now a viable drop-in replacement for $k$-means, and that GMM responsibilities can be leveraged to assign border vectors to multiple clusters. Our approach reaches fixed recall targets with up to $1.7\times$ fewer distance computations, or equivalently, yields $+2$--$12$ recall@10 at matched computational cost. We release the kernel as an open-source project.

📄 PDF Abstract BibTeX arXiv:2606.10896

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Fast Gauss Sums via Flash Attention

2026-09-04 · Nicolaj Rux, Sebastian Neumayer arxiv

Gaussian kernel sums are the computational core of maximum mean discrepancies (MMDs), kernel gradient flows, Stein variational gradient descent (SVGD), and many other kernel methods. At the same time, softmax attention h…

A Case Study in CUDA Kernel Fusion: Implementing FlashAttention-2 on NVIDIA Hopper Architecture using the CUTLASS Library

2023-12-19 · Ganesh Bikshandi, Jay Shah

We provide an optimized implementation of the forward pass of FlashAttention-2, a popular memory-aware scaled dot-product attention algorithm, as a custom fused CUDA kernel targeting NVIDIA Hopper architecture and writte…

GPU

SOCKET: SOft Collision Kernel EsTimator for Sparse Attention

2026-02-06 · Sahil Joshi, Agniva Chowdhury, Wyatt Bellinger, Amar Kanakamedala 외 arxiv

Exploiting sparsity during long-context inference is key to scaling large language models, as attention dominates the cost of autoregressive decoding. Sparse attention reduces this cost by restricting computation to a su…

Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regression

2025-10-01 · Yifei Zuo, Yutong Yin, Zhichen Zeng, Ang Li 외 arxiv

Transformer architectures have achieved remarkable success in various domains. While efficient alternatives to Softmax Attention have been widely studied, the search for more expressive mechanisms grounded in theoretical…

QFlash: Bridging Quantization and Memory Efficiency in Vision Transformer Attention

2026-04-28 · Sehyeon Oh, Yongin Kwon, Jemin Lee arxiv

FlashAttention improves efficiency through tiling, but its online softmax still relies on floating-point arithmetic for numerical stability, making full quantization difficult. We identify three main obstacles to integer…