Gopesh Bhardwaj
c798a92f36
* refactor docs
* Fixing docs build
* merging 967
* Improving readability with table usage
* review changes
* compilation fix
* README links
---------
Co-authored-by: srawat <120587655+SwRaw@users.noreply.github.com>
[ROCm/rocprofiler-sdk commit: 785cc39c16]
1.3 KiB
1.3 KiB
| 1 | Name | Calls | TotalDurationNs | AverageNs | Percentage | MinNs | MaxNs | StdDev |
|---|---|---|---|---|---|---|---|---|
| 2 | hipStreamCreateWithFlags | 4 | 262497406 | 65624351.500000 | 85.15 | 3991286 | 249121840 | 122332531.343496 |
| 3 | hipGetDeviceCount | 1 | 32505687 | 32505687.000000 | 10.54 | 32505687 | 32505687 | 0.00000000e+00 |
| 4 | hipHostMalloc | 12 | 6096409 | 508034.083333 | 1.98 | 443793 | 548024 | 39236.753678 |
| 5 | hipFree | 12 | 1994421 | 166201.750000 | 0.6470 | 7790 | 1036046 | 299086.860470 |
| 6 | hipMemcpyAsync | 12 | 1368378 | 114031.500000 | 0.4439 | 2490 | 764044 | 249308.051619 |
| 7 | hipMallocAsync | 12 | 927255 | 77271.250000 | 0.3008 | 51540 | 107671 | 20487.475966 |
| 8 | hipStreamSynchronize | 12 | 870486 | 72540.500000 | 0.2824 | 140 | 866606 | 250065.900069 |
| 9 | hipLaunchKernel | 16 | 692734 | 43295.875000 | 0.2247 | 1000 | 670044 | 167133.656647 |
| 10 | hipStreamDestroy | 4 | 619905 | 154976.250000 | 0.2011 | 92901 | 339252 | 122852.320356 |
| 11 | hipDeviceSynchronize | 4 | 404252 | 101063.000000 | 0.1311 | 570 | 385212 | 189518.505401 |
| 12 | hipHostFree | 12 | 271202 | 22600.166667 | 0.0880 | 11950 | 34950 | 7480.268600 |
| 13 | __hipRegisterFatBinary | 1 | 9000 | 9000.000000 | 2.920e-03 | 9000 | 9000 | 0.00000000e+00 |
| 14 | __hipRegisterFunction | 4 | 6150 | 1537.500000 | 1.995e-03 | 230 | 5370 | 2555.091323 |
| 15 | __hipPushCallConfiguration | 16 | 2460 | 153.750000 | 7.980e-04 | 70 | 1140 | 267.503894 |
| 16 | __hipPopCallConfiguration | 16 | 2000 | 125.000000 | 6.488e-04 | 70 | 680 | 151.613544 |
| 17 | hipGetLastError | 16 | 1270 | 79.375000 | 4.120e-04 | 50 | 440 | 96.295985 |
| 18 | hipSetDevice | 1 | 660 | 660.000000 | 2.141e-04 | 660 | 660 | 0.00000000e+00 |