Well, the problem is that if I do that, I'll lose all the extra calculation power HT gives me, and it will slow down a lot more the folding@home instances (which run in a lower priority) since the system and any other program I use will have to run on a core already running folding@home, thus "competing" for the whole core instead of "competing" for the same instructions (and they don't use exactly the same instructions since they are different programs) and slowing down a bit the whole thing.
And on the other hand, if I disable HT, Windows name the cores with numbers from 0 to 3 instead of the 0 to 7 needed for 8 cores, but I don't know if the 4 first ones are material and match with the 4 first ones of the 8 seen by windows when HT is enabled.
I'm quite sure that it does make a difference with everlasting folding@home work units but it'd be hard to guess...
I'll try to affix the same programs to some cores and find out (using super pi for example) and let you know.
Thank you, it will help me to find out !!
Nah, I think Windows thinks of every core, virtual or not, as a whole CPU and makes no distinction. I think it is not the CPU that determines itself its core load. Is it ?
I just ran my test.
Processors used (as seen in windows, set with the task manager) Time for performing an 1M PI calculation with Super Pi (seconds) 0 11 1 11 2 11 3 11 0+1 16 1+2 11 2+3 15 3+0 11
So cores 0 and 1 are the same material core. one is the actual one and one is the HT virtual one.
Same for cores 2 and 3.
Pattern is as follows : actual core, its associated virtual core, then actual core, its associated virtual core, and so on.
Material Core in Windows's taskmgr Virtual (HT) Core in Windows's taskmgr, associated with the material one on its left 0 1 2 3 4 5 6 7