As far as from what i've seen, HT doesn't make much of a difference in games. Toms hardware did an article on it iirc.
6 core vs 4 core gaming.
- Thread starter Master&Puppet
- Start date
As far as from what i've seen, HT doesn't make much of a difference in games. Toms hardware did an article on it iirc.
yassarikhan786
New member
Yeh, I read that article as well as others showing very little or no difference with HT on/off when gaming.
Rastalovich
New member
These are great tests to do. +rep on that score.
Some things to keep in mind. Many games have a "this is as fast as it's going to get" fps. Mostly due to the setups being well beyond the performance levels required. The matter that frames get to a certain level and then top out generally point to this. You could stick in 4x7970 or a dual 8 core xeon and they'd do the same.
Secondly, and this is a Windows thing I'm sceptical about, I personally believe when you change the cpu (i.e. change core numbers or threads available) ideally Windows needs re-installing. Sure it'll work well when you just make the change, but I don't think it works optimum.
(.. and of course you wait so many minutes for Windows to die down after booting, nominally 5 mins these days, to get the best cpu attention and it's not doing background Windows stuff. Make sure sppsvc.exe isn't running and using %age of cpu .. blah blah)
All depending on how accurate you wanna base your results. If it's just for fun and you do the same for all - so what.
Some things to keep in mind. Many games have a "this is as fast as it's going to get" fps. Mostly due to the setups being well beyond the performance levels required. The matter that frames get to a certain level and then top out generally point to this. You could stick in 4x7970 or a dual 8 core xeon and they'd do the same.
Secondly, and this is a Windows thing I'm sceptical about, I personally believe when you change the cpu (i.e. change core numbers or threads available) ideally Windows needs re-installing. Sure it'll work well when you just make the change, but I don't think it works optimum.
(.. and of course you wait so many minutes for Windows to die down after booting, nominally 5 mins these days, to get the best cpu attention and it's not doing background Windows stuff. Make sure sppsvc.exe isn't running and using %age of cpu .. blah blah)
All depending on how accurate you wanna base your results. If it's just for fun and you do the same for all - so what.
Master&Puppet
New member
Cheers
Yea, that would make the tests harder to do, I hadn't thought about that.
I saw that article too but I'm not very happy with the way in which the test was administered. Next time I do this (probs with a 3770k) I'll try and ensure a couple of other factors are addressed:
A. Make sure there isn't a 'ceiling' affecting the results i.e. a gpu bottleneck or frame rate ceiling as Rastalovich mentioned. The cpu has to be the limiting factor in the test.
B. It makes sense to me to try and ensure that the cpu cores are put under max load - if they aren't then who's to say that the system is going to shove processes onto the spare cores/threads when there is spare capacity already available.
Both of these issues seem to me to be solvable by heavy underclocking of the cpu.
At this stage I still believe that games are coded to use 2 or 4 cores (6 in the case of bf3) and that is that as far as game performance goes...however if your cpu begins maxing out, which I know is highly unlikely for most of us, then having the extra threads allows the system to offload non-game processes to the 5th-8th cores/threads thereby indirectly increasing game performance simply by freeing up capacity on the 4 threads being used by the game.
Interesting, I hadn't heard this one before - although as a matter of course I happen to re-install anyway, just because I like a clean install!
M&P
