| 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 : 236 |
| 6 |
Clock frequency : 1052 MHz |
| 7 |
|
| 8 |
Global memory bandwidth (GBPS) |
| 9 |
float : 62.52 |
| 10 |
float2 : 44.56 |
| 11 |
float4 : 76.55 |
| 12 |
float8 : 84.92 |
| 13 |
float16 : 2.15 |
| 14 |
|
| 15 |
Single-precision compute (GFLOPS) |
| 16 |
float : 1778.74 |
| 17 |
float2 : 1889.33 |
| 18 |
float4 : 1884.25 |
| 19 |
float8 : 1877.49 |
| 20 |
float16 : 1850.36 |
| 21 |
|
| 22 |
No half precision support! Skipped |
| 23 |
|
| 24 |
Double-precision compute (GFLOPS) |
| 25 |
double : 967.75 |
| 26 |
double2 : 966.69 |
| 27 |
double4 : 964.23 |
| 28 |
double8 : 958.01 |
| 29 |
double16 : 295.92 |
| 30 |
|
| 31 |
Integer compute (GIOPS) |
| 32 |
int : 968.24 |
| 33 |
int2 : 970.23 |
| 34 |
int4 : 968.07 |
| 35 |
int8 : 968.20 |
| 36 |
int16 : 958.80 |
| 37 |
|
| 38 |
Integer compute Fast 24bit (GIOPS) |
| 39 |
int : 968.37 |
| 40 |
int2 : 969.56 |
| 41 |
int4 : 967.91 |
| 42 |
int8 : 961.61 |
| 43 |
int16 : 950.62 |
| 44 |
|
| 45 |
Transfer bandwidth (GBPS) |
| 46 |
enqueueWriteBuffer : 1.86 |
| 47 |
enqueueReadBuffer : 3.45 |
| 48 |
enqueueWriteBuffer non-blocking : 3.34 |
| 49 |
enqueueReadBuffer non-blocking : 3.46 |
| 50 |
enqueueMapBuffer(for read) : 137.16 |
| 51 |
memcpy from mapped ptr : 3.02 |
| 52 |
enqueueUnmap(after write) : 6.91 |
| 53 |
memcpy to mapped ptr : 2.97 |
| 54 |
|
| 55 |
Kernel launch latency : 77.33 us |
| 56 |
|