Debunking the 100X GPU vs. CPU myth: an evaluation of throughput computing on CPU and GPU.
Victor W. LeeChangkyu KimJatin ChhuganiMichael DeisherDaehyun KimAnthony D. NguyenNadathur SatishMikhail SmelyanskiySrinivas ChennupatyPer HammarlundRonak SinghalPradeep DubeyPublished in: ISCA (2010)
Keyphrases
- graphics processing units
- graphics processors
- gpu implementation
- memory bandwidth
- heterogeneous computing
- general purpose
- graphics hardware
- intel xeon
- real time
- parallel computation
- parallel processing
- parallel computing
- parallel implementation
- floating point
- data transfer
- compute unified device architecture
- cpu implementation
- higher throughput
- real time rendering
- graphic processing unit
- multi threaded
- multithreading
- parallel programming
- level parallelism
- massively parallel
- processing power
- high performance computing
- general purpose computing
- response time
- cluster of workstations
- data sets
- compute intensive
- gpu accelerated
- efficient implementation
- evaluation method
- low latency