Post reply

Warning: this topic has not been posted in for at least 120 days.
Unless you're sure you want to reply, please consider starting a new topic.

Note: this post will not display until it's been approved by a moderator.

Message icon:


shortcuts: hit alt+s to submit/post or alt+p to preview

Topic Summary

Posted by: Admin
« on: April 22, 2014, 18:19:50 »

we change OpenCL behaviour and benchmark in future versions.
Lower values is better, and yes CPU speed shown only for 1 core, you need multiply it to 6 for true power of CPU. Similarity uses all cores than working you can see that in Task Manager.
Posted by: soylentgreen
« on: April 22, 2014, 14:24:51 »

When running the benchmark, it always says 1 cpu core. I've got a 6 core CPU, so can Similarity use all 6 cores?

I have a Nvidia GeForce 660, and when I am benchmarking the different work group sizes for the GPU, I get inconsistent results. I select 64 and get 74ms, then I select 128 and get 94ms, but when I select 64 again I get 90+ms. It changes when I go back and forth between the workgroup sizes.

What is better- workgroup size of 64 at 74ms, or a workgroup size of 256 at 150ms? Does that mean it's processing 3x more data and only 2x slower? (The CPU always shows 600ms)

I think it would be more accurate if there was a speed indicator during the analysis (like the number of files processed per second). That would be a better benchmark, since I can try different settings and run the analysis to compare speeds.