No description

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
fpga	gpu: always allocate page-sized chunks, then use LinearAllocator	2018-07-20 16:46:55 +02:00