CUDA toolkit documentation v8.0
http://docs.nvidia.com/cuda/cuda-getting-started-guide-for-microsoft-windows/index.html#axzz4aQaLXxZw
at limit for instruction-level parallelism per clock-cycle
-more processor to run computer faster
modern GPU: -thousands of ALUs
– hundreds of processors
– tens of thousands of concurrent threads
GPU
-smaller, faster, less power, more on chip
CPU
– complex control hardware
flexibility + performance
expensive in terms of power
GPU
– simpler control hardware
more hw for computation
potentially more power efficient
more restrictive programming model
