Monster clocks! Btw that PP is 51015, might wanna switch to GMH on the 12X multi, numbers don't lie. ;)
Printable View
Monster clocks! Btw that PP is 51015, might wanna switch to GMH on the 12X multi, numbers don't lie. ;)
Thank you guys......:)Quote:
Originally Posted by Gautam
Actualy if you see at the picture, it was done with 1 CPU core (within msconfig and numproc=1).....It is SLOWER that when you have 2 CPUs.....Next time I'll try with them both.....;)
:eek:... Man your CPU must beat 9 sec. :D.
I have noticed this also, but I have only compared 4 cores with 1 core. Maybe I should try 2 and 3 cores...Quote:
Originally Posted by hipro5
I am a n00b when it comes to this kind of cooling but at this speed and that vcore, will there be a big difference in temps if running 1 or 2 cores when running 'idle'?
Also, is it possible to dedicate one core for all the processes running in the background? If so, how?
P.S. With that SP1M result, maybe this thread should be copied to 'Extreme Overclocking'??? :D
This SP 1M run should be about 0.1s faster IF I had used both cores.....I'll retry it though with both cores....;)Quote:
Originally Posted by blossa
and yes you could "dedicate" one core for apps and the other for the bench....:)
I run on air one core vs. 2 cores, and I can't get 0.1s difference no way, of course I'm different league of overclocker then Hipro5, I disable one core via windows: run -> msconfig -> boot.ini -> advanced options -> numproc set to 1 and then reboot, does it always turn off same core?Quote:
Originally Posted by hipro5
awesome result Hipro :toast:
p.s. my result on air http://img345.imageshack.us/my.php?image=12297ec9.gif
4GHz to 5.5GHz to get 0.1s less is quite a difference, isn't it......?.I mean that you can't get 0.1s at 4GHz....You would propably get 0.02 - 0.04s less......:)Quote:
Originally Posted by Misko82
Maybe I didn't clearly said what I was thinking. My Spi 1M result fluctuate very very little 1 core vs. 2 cores, its like 0.04, everything else same, so when u go under 9.x sec if u are looking by %, 0.1 is bigger jump then on 12.x sec, so I'm asking how did u achieve so huge fluctuation, ok my oc knowledge isn't so impresive like yours, but where's the catch, of course I do tweaking, copywaza, maxmem etc...Quote:
Originally Posted by hipro5
As of 1M you don't actually need nor copywaza or maxmem......all you need is tweaks.....:)Quote:
Originally Posted by Misko82
At 5370MHz I have a 9.454s I think, so at 5550MHz, I could propably get a 9.100 or even lower....
I agree with that. ;)Quote:
Originally Posted by PallMall
Is it possible to set priorites so that they remain the same after reboot? I mean, maybe it will be an advantage to lower priority for some services...(?) And maybe even set it higher for some? ;)
for most system services you cannot set either afinity nor priority, all other services/apps should be disabledQuote:
Originally Posted by blossa
Way then I have lower times with maxmem and copywaza? Never mind is there some other tweak that I should be doing?Quote:
Originally Posted by hipro5
And why do u have 0.1sec difference when:
0.04:12.2=y:9.2
y=(0.04x9.2)/12.2
so u should be expecting something like 0.03, :confused:
You will probably see better times with them IF you are NOT using "full loaded" Windows with their effe's etc......At 4000MHz you MUST have a score BELOW that 13s......;)Quote:
Originally Posted by Misko82
What about using multi 12? Would give you slightly better score.
NOPE......A bit lower score......with those timmings I get about 35.6ns latency if I remember well.....;)Quote:
Originally Posted by NiCKE^
How can higher fsb and ram speed give you lower score :confused:
for higher FSB he should change RAM, he is using Fattys that aren't that good with high FSB as low latencys, if he would be using GMH or GKX would be a different story
But fatbodys handle volts better then GMH for example and I thought he wasn't affraid to give them a little volts.
Those Fatbodies can handle 2.8Vdimm max as for SP 1M and 2.74Vdimm as for the 23M......I don't want to work them at 4-3-2-1 and higher frequency which actualy can do about 493MHz at Cas 4-3-2-1.....;)
I've ALSO tested CellShok at 3-3-3-1 and x12 multi with no better resaults.......EVEN 32M with the fsb and the timmings I'm working right now IS FAST.....;)
Wow i didnt realize how huge the 3-2-2-1 boost was. What do you use to measure latency? Everest?Quote:
Originally Posted by hipro5
Yes Everest.....mostly and a bit S-Sandra.....:)
EDIT: Take a look at those two screenies.....PLUS on the second one the Rams ARE NOT maxed out.....;)
http://www.thelab.gr/images/Hipro5/C...442MHz_1_1.png
http://www.thelab.gr/images/Hipro5/C...408MHz_1_1.png
you are wrong, fatys up to ~2.7v, GMH 3.1 fairly easyQuote:
Originally Posted by NiCKE^
here is kunaak's explanation and I think hipro can confirm too
First to go sub 9s, never.Quote:
Originally Posted by PallMall
(You'll get a lot of reactions to this post, please keep that discussion out of this thread. It's Hipro's work-in-progress thread, not a new conspiricy theory thread !)
Quote:
Originally Posted by PallMall
Erm, first person in the world?
http://www.ocxtreme.org/onepagebook/...906spi1mwr.jpg
:confused:
Oh 2x512mb, got any improvement from 2gb on the times?