Standardizing GPU Radiation Test Approaches

worth of memory space, do nothing with it and then release the memory space. This keeps the process resident in the GPU pipeline for no less than 30 seconds per interval. The mutex can be adjusted to increase this by one to two orders of magnitude. The mutex memory allocation can only be performed in a serial fashion. Therefore, while all the computation calls can take place in parallel, each and every thread must wait in line to allocate the single integer space.