| 1 |
Platform: Intel(R) OpenCL |
| 2 |
Device: Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz |
| 3 |
Driver version : 1.2.0.43 (Linux x64) |
| 4 |
Compute units : 32 |
| 5 |
Clock frequency : 2600 MHz |
| 6 |
|
| 7 |
Global memory bandwidth (GBPS) |
| 8 |
float : 42.02 |
| 9 |
float2 : 41.93 |
| 10 |
float4 : 43.30 |
| 11 |
float8 : 42.25 |
| 12 |
float16 : 40.88 |
| 13 |
|
| 14 |
Single-precision compute (GFLOPS) |
| 15 |
float : 95.96 |
| 16 |
float2 : 382.23 |
| 17 |
float4 : 672.15 |
| 18 |
float8 : 192.14 |
| 19 |
float16 : 365.97 |
| 20 |
|
| 21 |
Double-precision compute (GFLOPS) |
| 22 |
double : 95.95 |
| 23 |
double2 : 349.98 |
| 24 |
double4 : 149.61 |
| 25 |
double8 : 189.77 |
| 26 |
double16 : 325.56 |
| 27 |
|
| 28 |
Transfer bandwidth (GBPS) |
| 29 |
enqueueWriteBuffer : 3.62 |
| 30 |
enqueueReadBuffer : 2.78 |
| 31 |
enqueueMapBuffer(for read) : 6523.34 |
| 32 |
memcpy from mapped ptr : 2.76 |
| 33 |
enqueueUnmap(after write) : 8815.61 |
| 34 |
memcpy to mapped ptr : 6.70 |
| 35 |
|
| 36 |
Kernel launch latency : 8.59 us |