AsmDB: Understanding and Mitigating Front-End Stalls in Warehouse-Scale Computers [abstract] (IEEE Xplore, Original Full Paper, PDF)
Nayana P. Nagendra, Grant Ayers, David I. August, Hyoun K. Cho, Svilen Kanev, Christos Kozyrakis, Trivikram Krishnamurthy, Heiner Litz, Tipp Moseley, and Parthasarathy Ranganathan
IEEE Micro, Volume 40, Number 3, May 2020.
IEEE Micro's "Top Picks" special issue "based on novelty and potential for long-term impact in the field of computer architecture" in 2019.
It is well known that the datacenters hosting today's cloud services waste
a significant number of cycles on front-end stalls. However, prior work has provided little
insights about the source of these front-end stalls and how to address them. This work
analyzes the cause of instruction cache misses at a fleet-wide scale and proposes a new
compiler-driven software code prefetching strategy to reduce instruction caches misses by 90%.