Quote from: spronkin1 on April 11, 2017, 03:28:31 PM

Quote from: joblo on April 09, 2017, 08:07:38 PM

Quote from: spronkin1 on April 09, 2017, 07:15:16 PM

I tried "-t 12 --cpu-affinity 0x555" (0b010101010101) and got a more high and stable result:
max 2300H/s, avg 2100 H/s on unlocked 2678v3

That's strange you only had 6 bits in the mask so you're running 2 threads on each CPU.
Try 0x555555. On my 6700K I get about the same with 4 of 8 threads with default mask or 0x55.

On each core or on each processor? I only have one Intel Xeon processor.
With the value 0x555555 I got 1220 H / s

I have no explanation for your results with a Xeon. My results differ.

The code will use the mask as a set of available logical CPUs to choose from. If there are more threads
than avaialable CPUs it is supposed to assign multiple threads per CPU.

Your results contradict that logic.

If you have the ability to monitor CPU usage on a per thread basis it may help to understand what's going on.
I'm not familiar with what tools are available on Windows but on Linux I can monotor CPU usage and core
temprerature. CPU usage monitors the logical CPUs while core temperature indicates how hard the physical
cores (hyperthreaded) are working.