Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Locality-Driven Dynamic GPU Cache Bypassing

Conference ·
This paper presents novel cache optimizations for massively parallel, throughput-oriented architectures like GPUs. Based on the reuse characteristics of GPU workloads, we propose a design that integrates such efficient locality filtering capability into the decoupled tag store of the existing L1 D-cache through simple and cost-effective hardware extensions.
Research Organization:
Pacific Northwest National Laboratory (PNNL), Richland, WA (US)
Sponsoring Organization:
USDOE
DOE Contract Number:
AC05-76RL01830
OSTI ID:
1194296
Report Number(s):
PNNL-SA-109271; KJ0402000
Country of Publication:
United States
Language:
English

Similar Records

RACB: Resource Aware Cache Bypass on GPUs
Conference · Wed Oct 01 00:00:00 EDT 2014 · 2014 International Symposium on Computer Architecture and High Performance Computing Workshop; 22-24 Oct. 2014; Paris, France · OSTI ID:1567596

A performance model for GPUs with caches
Journal Article · Mon Jun 23 20:00:00 EDT 2014 · IEEE Transactions on Parallel and Distributed Systems · OSTI ID:1333005

Locality-Aware CTA Clustering For Modern GPUs
Conference · Sat Apr 08 00:00:00 EDT 2017 · OSTI ID:1355097