Daniel Krebs 8a06e96e92 gpu: always allocate page-sized chunks, then use LinearAllocator ... This was neccessary in order to make the memory available via GDRcopy when multiple small allocations were made. cudaMalloc() would return multiple memory chunks located in the same GPU page, which GDRcopy pretty much dislikes (`gdrdrv:offset != 0 is not supported`). As a side effect, this will keep the number of BAR-mappings done via GDRcopy low, because they seem to be quite limited.		2018-07-20 16:46:55 +02:00
..
gdrcopy@0441daa447	gpu: implement basic GPU plugin that can do DMA to and from its memory	2018-05-15 18:15:17 +02:00
include/villas	gpu: always allocate page-sized chunks, then use LinearAllocator	2018-07-20 16:46:55 +02:00
CMakeLists.txt	gpu: implement basic GPU plugin that can do DMA to and from its memory	2018-05-15 18:15:17 +02:00
gpu.cpp	gpu: always allocate page-sized chunks, then use LinearAllocator	2018-07-20 16:46:55 +02:00
kernels.cu	gpu: implement basic GPU plugin that can do DMA to and from its memory	2018-05-15 18:15:17 +02:00
kernels.hpp	gpu: implement basic GPU plugin that can do DMA to and from its memory	2018-05-15 18:15:17 +02:00