| 1 |
|
| 2 |
Platform: Intel(R) OpenCL |
| 3 |
Device: Intel(R) Many Integrated Core Acceleration Card |
| 4 |
Driver version : 1.2 (Linux x64) |
| 5 |
Compute units : 240 |
| 6 |
Clock frequency : 1333 MHz |
| 7 |
|
| 8 |
Global memory bandwidth (GBPS) |
| 9 |
float : 68.00 |
| 10 |
float2 : 49.67 |
| 11 |
float4 : 89.31 |
| 12 |
float8 : 101.30 |
| 13 |
float16 : 1.84 |
| 14 |
|
| 15 |
Single-precision compute (GFLOPS) |
| 16 |
float : 2130.58 |
| 17 |
float2 : 2262.12 |
| 18 |
float4 : 2255.07 |
| 19 |
float8 : 2247.39 |
| 20 |
float16 : 2213.89 |
| 21 |
|
| 22 |
No half precision support! Skipped |
| 23 |
|
| 24 |
Double-precision compute (GFLOPS) |
| 25 |
double : 1157.62 |
| 26 |
double2 : 1156.44 |
| 27 |
double4 : 1153.55 |
| 28 |
double8 : 1146.15 |
| 29 |
double16 : 354.26 |
| 30 |
|
| 31 |
Integer compute (GIOPS) |
| 32 |
int : 1158.72 |
| 33 |
int2 : 1161.01 |
| 34 |
int4 : 1158.69 |
| 35 |
int8 : 1158.58 |
| 36 |
int16 : 1147.02 |
| 37 |
|
| 38 |
Integer compute Fast 24bit (GIOPS) |
| 39 |
int : 1158.82 |
| 40 |
int2 : 1160.21 |
| 41 |
int4 : 1158.02 |
| 42 |
int8 : 1150.32 |
| 43 |
int16 : 1136.05 |
| 44 |
|
| 45 |
Transfer bandwidth (GBPS) |
| 46 |
enqueueWriteBuffer : 2.25 |
| 47 |
enqueueReadBuffer : 5.62 |
| 48 |
enqueueWriteBuffer non-blocking : 4.92 |
| 49 |
enqueueReadBuffer non-blocking : 5.57 |
| 50 |
enqueueMapBuffer(for read) : 137.80 |
| 51 |
memcpy from mapped ptr : 3.71 |
| 52 |
enqueueUnmap(after write) : 6.55 |
| 53 |
memcpy to mapped ptr : 3.43 |
| 54 |
|
| 55 |
Kernel launch latency : 79.60 us |
| 56 |
|
| 57 |
|
| 58 |
Platform: NVIDIA CUDA |
| 59 |
|
| 60 |
Platform: AMD Accelerated Parallel Processing |