Does anyone know if allocating more threads for Commerical detection signicantly impacts the commercial detection routine? If so does anyone know where the dimenishing return is? Would be beneficial to allocate only 1 thread or all possible threads?
Here is the same 1 hour h264 TVE tv show
1 thread: finished in 16m38.476242275s
2 threads: finished in 9m35.732427208s
5 threads: finished in 4m39.551914119s
7 threads: finished in 4m33.719244678s
10 threads: finished in 3m32.847742912s
All tests found the same amount of markers.
CPU: i5-10400 16GB ram
Based on these numbers it looks like the sweet spot is a cpu with a passmark of around 6000. I don’t currently have any OTA MPEG’s 2 recordings to test it on. Maybe later I’ll test one out.
1 Like
Looks like the SweetSpot is 5 threads, after that the time used was negligible till it got up to 10 threads.
I'm at a loss to understand how the benchmark of 6k was determined by the info here. His listed CPU benchmark was 12+k and a single thread is listed at ~2600 I believe that CPU is 6 processors with hyper-threading making it look like 12 CPUs.
-Bill
Just an unscientific observation. As you say it’s a 12k passmark with 6/12 threads so I took an estimated 1k per thread and added a thread for os. So 5 for comskip and one for OS is 6 at 1000 each is how I got a 6k passmark. Feel free to run some tests with your setup and I’d gladly edit my post if it turns out to be completely off base.
1 Like
Got it.
1 Like
The answer will depend on your CPU. In my experimenting with faster CPUs by Intel and AMD, 2 threads is quick and 3 a little quicker. With a low end CPU, 1 or 2 will get the best results.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.