Close

Presentation

A Near-Storage Framework for Boosted Data Preprocessing of Mass Spectrum Clustering
TimeTuesday, July 12th3:30pm - 3:50pm PDT
Location3005, Level 3
Event Type
Research Manuscript
Keywords
SoC, Heterogeneous, and Reconfigurable Architectures
Topics
Design
DescriptionWe propose MSAS to speed up Mass spectrometry(MS) preprocessing. Instead of loading data into host memory and CPU, MSAS processes spectra near storage, thus reducing the expensive cost of data movement. We present three types of accelerators that leverage internal bandwidth at three different storage levels. Our results show the channel-level design yields the best performance improvement for preprocessing - it is up to 187x faster than running on CPU. After integrating into existing MS clustering tools, it achieves 3.5x speedup with 2.8x better energy efficiency.