摘要
系统地探讨使用CPU/GPU协同处理理论对高分辨率卫星影像进行MTF补偿的方法。首先在GPU上对方法进行基本实现,并通过3种性能优化策略(执行配置优化、存储访问优化和指令优化)进一步提高了方法的执行效率。在Intel Xeon E5650CPU和NVIDIA Tesla C2050GPU组成的CPU/GPU系统中对高分一号卫星全色影像进行MTF补偿,加速比达到42.80倍。在此基础上,为充分利用CPU的计算性能,使用CPU/GPU负载分配策略将部分负载分配给CPU进行处理,使用该策略后,方法加速比达到47.82倍,相应的处理时间压缩至1.62s,可满足对高分辨率卫星影像进行近实时MTF补偿的需求。
A novel workload-distribution based CPU/GPU MTF compensation approach for high resolution satellite images is proposed.First,the basic GPU implementation issues are addressed;next,three performance tuning methods,the execution configuration optimization,memory access optimization and instruction optimization,are applied to further improve the performance.The approach is tested with the GF-1panchromatic image in the CPU/GPU system that consists of an Intel Xeon E5650 CPU and a NVIDIA Tesla C2050 GPU.The experimental result shows that the speedup ratio is up to 42.80 times.Furthermore,the CPU/GPU workload distribution strategy is presented to fully exploit CPU's computing horsepower.With this strategy,the speedup ratio of MTF compensation finally reaches to 47.82times(the corresponding processing time is 1.62s),which could meet the requirement of near real-time MTF compensation for high resolution satellite images.
出处
《测绘学报》
EI
CSCD
北大核心
2014年第6期598-606,共9页
Acta Geodaetica et Cartographica Sinica
基金
国家973计划(2014CB744201
2012CB719902
2012CB719901)
国家863计划(2011AA120203)
中央高校基本科研业务费专项资金(2012619020205)
长江学者与创新团队发展计划(IRT1278)
国家自然科学基金(41371430)
关键词
CPU/GPU
调制传递函数补偿
执行配置优化
存储访问优化
指令优化
负载分配策略
CPU/GPU
MTF compensation
execution configuration optimization
memory access optimization
instruction optimization
workload distribution strategy
