Abstract | ||
---|---|---|
Graphics processing units (GPUs) are becoming an increasingly popular platform to run applications that require a high computation throughput. They are limited, however, by memory bandwidth and power and, as such, cannot always achieve their full potential. This paper presents the PUMA architecture - a domain-specific accelerator designed specifically for medical imaging applications, but with sufficient generality to make it programmable. The goal is to closely match the performance achieved by GPUs in this domain but at a fraction of the power consumption. The results are quite promising - PUMA achieves upto 2X the performance of a modern GPU architecture and has upto a 54X improved efficiency on a floating-point and memory-intensive MRI reconstruction algorithm. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1109/SASP.2009.5226332 | 2009 IEEE 7TH SYMPOSIUM ON APPLICATION SPECIFIC PROCESSORS (SASP 2009) |
Keywords | Field | DocType |
i. i ntroduction,memory bandwidth,parallel processing,computer architecture,graphics,acceleration,benchmark testing,power efficiency,image analysis,registers,hardware,bandwidth,coprocessors,floating point,magnetic resonance imaging | Iterative reconstruction,Graphics,Memory bandwidth,Computer science,CUDA,Parallel computing,Image processing,Coprocessor,Throughput,Computer hardware,Benchmark (computing) | Conference |
Citations | PageRank | References |
2 | 0.38 | 17 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ganesh S. Dasika | 1 | 387 | 24.30 |
Kevin Fan | 2 | 335 | 20.29 |
Scott Mahlke | 3 | 4811 | 312.08 |