cuda.h defines CUdeviceptr as an unsigned int. Is it possible to allocate more the 4GB using the driver api? Fermi supports 64-bit pointers and there are errors for improperly using 64-bit pointers (CUDA_ERROR_POINTER_IS_64BIT and CUDA_ERROR_SIZE_IS_64BIT). Also, adding integers to a CUdeviceptr seems to produce new valid CUdeviceptrs at the appropriate offset from the original pointer, is this behavior specified in the manual somewhere, or could it change in future implementations? Thanks.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| CUdeviceptr is wrong size on 64-bit platforms 64-bit, CUdeviceptr, Linux | 0 | 5490 | October 16, 2008 | |
| 32b / 64b question - CUdeviceptr size | 5 | 18927 | February 5, 2009 | |
| CUdeviceptr should be typdedef'd as void *, instead of unsigned int | 3 | 6970 | June 2, 2010 | |
| [bugreport] __alignof(CUdeviceptr) == 4, should 8 | 12 | 27203 | July 5, 2010 | |
| Are 64-bit integer instructions natively supported by GPU? | 1 | 2318 | October 5, 2009 | |
| CUDA pointers: 64-bit or not? | 2 | 3406 | June 19, 2008 | |
| Fermi and zero-copy | 0 | 1171 | November 20, 2009 | |
| Fermi (2.0) cuda device on 64-bit Linux with 32bit device code | 3 | 10303 | February 13, 2011 | |
| cudaHostAlloc 32-bit pointer limit | 2 | 1006 | August 30, 2010 | |
| 64 bits device code | 4 | 6185 | September 2, 2010 |