TIME thru                            Starting, CPU:  0.33 (d= 0.33)
Thrust v1.7, CUDA_VERSION: 6,000, CUDA_ARCH: 0
N=500,000,000
Comparing sort on host and device:
TIME thru              Create int vec on host, CPU:  1.06 (d= 0.73)
TIME thru       Transf random int vec on host, CPU:  1.82 (d= 0.76)
TIME thru                        Sort on host, CPU: 12.98 (d=11.16)
Before creating device vector: cuda_mem_tot=6,039,339,008, cuda_mem_free=5,951,418,368
TIME thru            Create int vec on device, CPU: 12.98 (d= 0.00)
After creating device vector: cuda_mem_tot=6,039,339,008, cuda_mem_free=3,951,276,032
TIME thru     Transf random int vec on device, CPU: 13.00 (d= 0.02)
TIME thru                      Sort on device, CPU: 13.65 (d= 0.65)
TIME thru           Transferring back to host, CPU: 14.17 (d= 0.52)
Comparing reduce on host and device:
TIME thru            Create float vec on host, CPU: 14.90 (d= 0.73)
TIME thru     Transf random float vec on host, CPU: 15.73 (d= 0.83)
TIME thru                      Reduce on host, CPU: 15.73 (d= 0.00)
TIME thru          Create float vec on device, CPU: 15.74 (d= 0.01)
TIME thru   Transf random float vec on device, CPU: 15.75 (d= 0.01)
TIME thru                    Reduce on device, CPU: 15.77 (d= 0.02)