-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Online CPU Speed Test
- Thread starter ben9997
- Start date
- Joined
- Sep 29, 2004
I'd wair for bechmarks before saying that.
Benchies I've seen show Sandy bridge being roughly 10% faster clock for clock to their current gen stuff. Bulldozer is a radically different architecture to anything out today so it's definitely going to be interesting to see how it performs.
I believe we're going to be back to the old days of AMD being faster for some tasks and Intel faster on others. If thats the case, then choice of CPU will really boild down to what you use the system for and how much you're willing to spend.
- Joined
- Jun 23, 2012
Holy DEAD THREAD revival.
I know I put my system info down & it came out completly diffrent from what I put in! I had put down windows 7 and other specs. I don't know what is worse the Internet or the PC
New technology A!
- Joined
- May 10, 2009
Strange test, only uses three threads as far as I can tell. It hits the follow score on my chip at ~4.3GHz and will not go higher, even if I crank it to 4670MHz. It completely caps out here.
(Yeah, EIST, should be a 45x multi)
It's location sensing bits are hilarious, too.
Tragic really as given the scaling from 17.xx at 3.7GHz to 23.86 at 4.2-4.3GHz I was on track to take over that top spot
I do love me some benching competition
I'll have to run this on my BD when I get a chance.
(Yeah, EIST, should be a 45x multi)
It's location sensing bits are hilarious, too.
Tragic really as given the scaling from 17.xx at 3.7GHz to 23.86 at 4.2-4.3GHz I was on track to take over that top spot
I do love me some benching competition
I'll have to run this on my BD when I get a chance.
- Joined
- Jun 23, 2012
I got you all beat!
Do note the fact I am Folding while running the test.
Do note the fact I am Folding while running the test.
Strange test, only uses three threads as far as I can tell. It hits the follow score on my chip at ~4.3GHz and will not go higher, even if I crank it to 4670MHz. It completely caps out here.
View attachment 113868
(Yeah, EIST, should be a 45x multi)
It's location sensing bits are hilarious, too.
Tragic really as given the scaling from 17.xx at 3.7GHz to 23.86 at 4.2-4.3GHz I was on track to take over that top spot
I do love me some benching competition
I'll have to run this on my BD when I get a chance.
Just throwing a wild idea out there..... are your cores parked? mine only ran a slower test (about 24) first time round, so i ran it again and again and suddenly it seemed like something came unstuck and it ran the higher score over and over again.
Win 7 i think is a bit slow to unpark cores.
http://www.overclockers.com/forums/showthread.php?t=713005
I think what its doing is reading the registry, its reading what Win 7 is set to park idle cores, not the ones which are parked at the time, if you unpark the cores with it it rewrites the registry to tell the OS to unpark them even when idle.
I came a cross it while searching for an answer as to why it is that when ever i bench from idle cores i have to run the bench over and over in a row a couple of times to get the score i'm expecting, it does it all the time.
Its as if i have a sticky core or something.
- Joined
- May 10, 2009
AMD chips are notoriously slow at coming out of CnQ, I'm assuming that it is intentional.
The following is conjecture and theory:
It makes sense really, most jobs that "require" a core to run 100% don't last very long at all, so the core can do it in say 10ms at full speed or 80ms at idle speed.
If the core goes "ZOMG I'MA BE USEFUL!" and slams itself to high speed it eats a bunch of power, and then sits there eating power before going back to idle.
If the core thinks about it and says "Well, let's see if this job actually needs that much power. Then I'll decide" it will plow through those tasks using a lot less power.
Of course if the task is a benchmark then every ms counts (ever ns if you're benching seriously) and the score isn't as high.
Intel CPUs seem to come out of EIST rather more quickly. Again I expect it's a conscious decision.
In any case I'll try disabling the parking with that app and rebooting and re-testing. Given the cap and how it acts frequency wise I don't expect it to change anything, but I'm curious now!
The following is conjecture and theory:
It makes sense really, most jobs that "require" a core to run 100% don't last very long at all, so the core can do it in say 10ms at full speed or 80ms at idle speed.
If the core goes "ZOMG I'MA BE USEFUL!" and slams itself to high speed it eats a bunch of power, and then sits there eating power before going back to idle.
If the core thinks about it and says "Well, let's see if this job actually needs that much power. Then I'll decide" it will plow through those tasks using a lot less power.
Of course if the task is a benchmark then every ms counts (ever ns if you're benching seriously) and the score isn't as high.
Intel CPUs seem to come out of EIST rather more quickly. Again I expect it's a conscious decision.
In any case I'll try disabling the parking with that app and rebooting and re-testing. Given the cap and how it acts frequency wise I don't expect it to change anything, but I'm curious now!
- Joined
- Oct 11, 2005
- Location
- Tau'ri
I will run it on my quad core BD later today.
Just ran this on my 1090T (stock). 1600 mem 2400 cpunb.
CnQ is enabled, and historically have had much better success with CnQ then I have with EIST.
EDIT: Ran the core parking app it says 3cores are parked. This is a 24/7 setup up though downclocking and power management are important. So I am not going to unpark anything. I will run it on the BD later thats a temporary setup, and already has the MS hotfix to fix parking and threading.
Just ran this on my 1090T (stock). 1600 mem 2400 cpunb.
CnQ is enabled, and historically have had much better success with CnQ then I have with EIST.
EDIT: Ran the core parking app it says 3cores are parked. This is a 24/7 setup up though downclocking and power management are important. So I am not going to unpark anything. I will run it on the BD later thats a temporary setup, and already has the MS hotfix to fix parking and threading.
Last edited:
- Joined
- Oct 11, 2005
- Location
- Tau'ri
^^^^ get that Thuban crancked up, man, come on..... 26.84 to beat
im on a dead silent setup. (noctua c12-d14, passive cooled 6770 and a coolermaster silencio case
No overclocking needed nor wanted for now. I will play around with it after the parts for my x79 get here
(6c/12t zigmatek elysium 580GTx and 1866 quad channel ram
) That wont be a 24/7 either but might serve that capacity for a couple of weeks 
