InvisiBill
2[H]4U
- Joined
- Jan 2, 2003
- Messages
- 2,608
There aren't a whole lot of posted dnetc benchmark results for the HD5000's, and the ones that are posted have some some pretty big variances. I trust you guys who know about hardware a little more than I trust the random results posted there. I'm specifically looking at the 5870, but I'd like to see other lower cards too for comparison.
You should end up with something like this (obviously this example is from the CUDA version):
- Go to the Pre-Release Clients page and download the [x86/Stream] client.
- Unzip the file somewhere.
- Run dnetc.exe -runoffline -l benchmark.txt -cpuinfo
- Run dnetc.exe -runoffline -l benchmark.txt -bench
- Post your results here. "CODE" tags will keep it a manageable size.
You should end up with something like this (obviously this example is from the CUDA version):
Code:
dnetc v2.9107-516-GTR-09122714 for CUDA 2.2 on Win32 (WindowsNT 6.1).
[Mar 31 15:40:29 UTC] nvcuda.dll Version: 8.17.11.9716
Automatic processor identification tag: 8192
name: Quadro NVS 160M (1 MPs)
Estimated processor clock speed (0 if unknown): 1450 MHz
Number of processors detected by this client: 1
Number of processors supported by this client: 128
dnetc v2.9107-516-GTR-09122714 for CUDA 2.2 on Win32 (WindowsNT 6.1).
[Mar 31 15:40:32 UTC] nvcuda.dll Version: 8.17.11.9716
[Mar 31 15:40:32 UTC] RC5-72: using core #0 (CUDA 1-pipe 64-thd).
[Mar 31 15:40:51 UTC] RC5-72: Benchmark for core #0 (CUDA 1-pipe 64-thd)
0.00:00:16.52 [6,410,666 keys/sec]
[Mar 31 15:40:51 UTC] RC5-72: using core #1 (CUDA 1-pipe 128-thd).
[Mar 31 15:41:10 UTC] RC5-72: Benchmark for core #1 (CUDA 1-pipe 128-thd)
0.00:00:16.38 [8,615,123 keys/sec]
[Mar 31 15:41:10 UTC] RC5-72: using core #2 (CUDA 1-pipe 256-thd).
[Mar 31 15:41:32 UTC] RC5-72: Benchmark for core #2 (CUDA 1-pipe 256-thd)
0.00:00:16.30 [9,823,934 keys/sec]
[Mar 31 15:41:32 UTC] RC5-72: using core #3 (CUDA 2-pipe 64-thd).
[Mar 31 15:41:52 UTC] RC5-72: Benchmark for core #3 (CUDA 2-pipe 64-thd)
0.00:00:17.75 [8,775,579 keys/sec]
[Mar 31 15:41:52 UTC] RC5-72: using core #4 (CUDA 2-pipe 128-thd).
[Mar 31 15:42:16 UTC] RC5-72: Benchmark for core #4 (CUDA 2-pipe 128-thd)
0.00:00:18.06 [8,881,913 keys/sec]
[Mar 31 15:42:16 UTC] RC5-72: using core #6 (CUDA 4-pipe 64-thd).
[Mar 31 15:42:40 UTC] RC5-72: Benchmark for core #6 (CUDA 4-pipe 64-thd)
0.00:00:19.26 [10,039,487 keys/sec]
[Mar 31 15:42:40 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Mar 31 15:42:43 UTC] RC5 CUDA ERROR [0]: 'the launch timed out and was terminated' (cudaEventSynchronize)
[Mar 31 15:42:44 UTC] RC5-72 benchmark summary :
Default core : #0 (CUDA 1-pipe 64-thd)
Fastest core : #6 (CUDA 4-pipe 64-thd)
[Mar 31 15:42:44 UTC] Core #6 is significantly faster than the default core.
The GPU core selection has been made as a tradeoff between core speed
and responsiveness of the graphical desktop.
Please file a bug report along with the output of -cpuinfo
only if the the faster core selection does not degrade graphics performance.