| 1 |
|
| 2 |
Platform: NVIDIA CUDA |
| 3 |
Device: GeForce GT 750M |
| 4 |
Driver version : 335.23 (Win32) |
| 5 |
Compute units : 2 |
| 6 |
Clock frequency : 1085 MHz |
| 7 |
|
| 8 |
Global memory bandwidth (GBPS) |
| 9 |
float : 24.51 |
| 10 |
float2 : 25.27 |
| 11 |
float4 : 25.68 |
| 12 |
float8 : 13.01 |
| 13 |
float16 : 12.34 |
| 14 |
|
| 15 |
Single-precision compute (GFLOPS) |
| 16 |
float : 580.07 |
| 17 |
float2 : 765.75 |
| 18 |
float4 : 739.55 |
| 19 |
float8 : 757.51 |
| 20 |
float16 : 740.08 |
| 21 |
|
| 22 |
Double-precision compute (GFLOPS) |
| 23 |
double : 37.16 |
| 24 |
double2 : 37.11 |
| 25 |
double4 : 37.05 |
| 26 |
double8 : 36.90 |
| 27 |
double16 : 36.61 |
| 28 |
|
| 29 |
Transfer bandwidth (GBPS) |
| 30 |
enqueueWriteBuffer : 2.52 |
| 31 |
enqueueReadBuffer : 2.39 |
| 32 |
enqueueMapBuffer(for read) : 2.48 |
| 33 |
memcpy from mapped ptr : 2.65 |
| 34 |
enqueueUnmap(after write) : 2.40 |
| 35 |
memcpy to mapped ptr : 2.79 |
| 36 |
|
| 37 |
Kernel launch latency : 100.45 us |
| 38 |
|