I am currently working on a Tesla C1060 under Windows XP 64. According to this thread: http://forums.nvidia.com/index.php?showtopic=98291&st=0, one could successfully allocate a maximum of ~4,232,800,000 bytes on the device. I’m wondering how I can calculate this number. When running cudaGetDeviceProperties(), I can obtain the device properties, but I can’t figure out where the 4,232,800,000 comes from. Or, is there a table published somewhere that indicates the maximum amount of memory that can be allocated for each device? Thanks in advance!