arXivDaily arXiv每日学术速递 周一至周五更新
看起来今天可以划水了
arXiv 2103.04808cs.ARcs.IR
收藏

基于FPGA近似嵌入相似性的Top-K SpMV的HBM效率提升

Scaling up HBM Efficiency of Top-K SpMV for Approximate Embedding Similarity on FPGAs

Alberto Parravicini, Luca Giuseppe Cellamare, Marco Siracusa, Marco Domenico Santambrogio

摘要:

Top-K SpMV是稀疏嵌入相似性搜索的关键组成部分。这种稀疏的工作负载在采用传统缓存策略的通用NUMA系统上无法很好地执行。相反,现代FPGA加速器卡有一些锦囊妙计。我们引入了Top-K SpMV FPGA设计,该设计利用了降低的精度和新颖的逐包CSR矩阵压缩,实现了自定义数据布局,并提供了即使在峰值带宽更高的架构中也无法实现的带宽效率。使用基于HBM的开发板,我们比多线程CPU实现快100倍,比GPU快2倍,带宽高20%,能效高14.2倍。

英文摘要:

Top-K SpMV is a key component of similarity-search on sparse embeddings. This sparse workload does not perform well on general-purpose NUMA systems that employ traditional caching strategies. Instead, modern FPGA accelerator cards have a few tricks up their sleeve. We introduce a Top-K SpMV FPGA design that leverages reduced precision and a novel packet-wise CSR matrix compression, enabling custom data layouts and delivering bandwidth efficiency often unreachable even in architectures with higher peak bandwidth. With HBM-based boards, we are 100x faster than a multi-threaded CPU implementation and 2x faster than a GPU with 20% higher bandwidth, with 14.2x higher power-efficiency.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑