Abstract | ||
---|---|---|
Memory performance is one essential factor for tapping into the full potential of the massive parallelism of GPU. It has motivated some recent efforts in GPU cache modeling. This paper presents a new data-centric way to model the performance of a system with heterogeneous memory resources. The new model is composable, meaning it can predict the performance difference due to placing data differently by profiling the execution just once. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1145/2851141.2851182 | PPOPP |
Keywords | Field | DocType |
Locality metrics,Locality modeling,Footprint | Database-centric architecture,Computer science,Cache,Massively parallel,Profiling (computer programming),Parallel computing,Theoretical computer science,Combinatorial optimization,Footprint | Conference |
Volume | Issue | ISSN |
51 | 8 | 0362-1340 |
Citations | PageRank | References |
3 | 0.37 | 4 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hao Luo | 1 | 19 | 2.25 |
Guoyang Chen | 2 | 112 | 5.87 |
Pengcheng Li | 3 | 28 | 3.78 |
Chen Ding | 4 | 749 | 43.96 |
Xipeng Shen | 5 | 2025 | 118.55 |