Abstract | ||
---|---|---|
Limited memory bandwidth and long access latency are serious GPU performance bottlenecks. For emerging applications that are inherently error resilient, this paper proposes to predict the values of safe-to-approximate loads and drop a certain fraction of the cache misses with predicted values to improve performance and energy efficiency. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1109/MDAT.2015.2504899 | IEEE Design & Test |
Keywords | Field | DocType |
Bandwidth,Graphics processing units,Memory management,Safety,Degradation,Energy efficiency | Bottleneck,Cache pollution,Latency (engineering),Computer science,Cache,Efficient energy use,Bandwidth (signal processing),Memory management,Cache coloring,Computer engineering | Journal |
Volume | Issue | ISSN |
33 | 1 | 2168-2356 |
Citations | PageRank | References |
8 | 0.42 | 12 |
Authors | ||
6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Amir Yazdanbakhsh | 1 | 241 | 15.28 |
Bradley Thwaites | 2 | 140 | 5.67 |
H. Esmaeilzadeh | 3 | 1443 | 69.71 |
Gennady Pekhimenko | 4 | 706 | 28.75 |
Onur Mutlu | 5 | 9446 | 357.40 |
Todd C. Mowry | 6 | 3021 | 253.75 |