| 1 |
|
| 2 |
Platform: NVIDIA CUDA |
| 3 |
Device: GeForce RTX 2060 |
| 4 |
Driver version : 440.33.01 (Linux x64) |
| 5 |
Compute units : 30 |
| 6 |
Clock frequency : 1200 MHz |
| 7 |
|
| 8 |
Global memory bandwidth (GBPS) |
| 9 |
float : 277.69 |
| 10 |
float2 : 292.44 |
| 11 |
float4 : 299.43 |
| 12 |
float8 : 307.26 |
| 13 |
float16 : 295.93 |
| 14 |
|
| 15 |
Single-precision compute (GFLOPS) |
| 16 |
float : 6660.58 |
| 17 |
float2 : 6652.10 |
| 18 |
float4 : 6616.57 |
| 19 |
float8 : 6490.34 |
| 20 |
float16 : 6415.59 |
| 21 |
|
| 22 |
No half precision support! Skipped |
| 23 |
|
| 24 |
Double-precision compute (GFLOPS) |
| 25 |
double : 225.49 |
| 26 |
double2 : 222.73 |
| 27 |
double4 : 223.70 |
| 28 |
double8 : 222.83 |
| 29 |
double16 : 220.67 |
| 30 |
|
| 31 |
Integer compute (GIOPS) |
| 32 |
int : 5287.57 |
| 33 |
int2 : 5311.86 |
| 34 |
int4 : 5324.80 |
| 35 |
int8 : 5347.13 |
| 36 |
int16 : 5331.60 |
| 37 |
|
| 38 |
Transfer bandwidth (GBPS) |
| 39 |
enqueueWriteBuffer : 6.02 |
| 40 |
enqueueReadBuffer : 6.20 |
| 41 |
enqueueMapBuffer(for read) : 5.98 |
| 42 |
memcpy from mapped ptr : 13.10 |
| 43 |
enqueueUnmap(after write) : 6.15 |
| 44 |
memcpy to mapped ptr : 13.22 |
| 45 |
|
| 46 |
Kernel launch latency : 3.44 us |
| 47 |
|