Overclocking-Team with Prescott-OC at 6.207 MHZ (326,7 x 19)
DeadLinkremoved
also anounced here tonight (post 1120 at the end of the page):
DeadLinkremoved
congratulations!
Printable View
Overclocking-Team with Prescott-OC at 6.207 MHZ (326,7 x 19)
DeadLinkremoved
also anounced here tonight (post 1120 at the end of the page):
DeadLinkremoved
congratulations!
1st link doesn't work for me. :(
Edit: DeadLinkremoved
Awesome stuff! Congrats
I can't read German - but was this the max benchable?: DeadLinkremoved
Another screen shot :rolleyes:
Am I missing the pictures of the benchmarks over 6ghz here? I couldn't find any in the 29page thread.
thanks
PS: I hope they can do more than 5.3ghz for benches with SS at 6.2ghz :stick: :stick:
Friggin links are down for me.
I can't understand that: i can see the sreenshots (and attachments) good and clear.
i try to enclose them with this post.
sorry so far for the problems, but theres 100% no cheating in that!
any way of modding the mobo so the processor gets some good clean steady voltage? or an expensive PSU? maybe can bench at 6ghz then.
where is the list of motherboard, ram ecc ecc used????
maybe not enough power/juice from current psu ??? 6.2ghz @ 1.75v is like 293W of power :eek:Quote:
Originally Posted by pc ice
yea.... mobo cant supply enough power.
what mobo was used??? i think fatal1ty right????
s, i think abit (and an dfi 875
with the dfi they also runned chipset under LN2....nice....
i need to know which abit is this....
It was a fatality AA8XE
Looks like AS8 to me...
http://xtremesystems.org/forums/atta...&stc=1&thumb=1
Look at the Ram, they have 2,25V, no ram with DDR1 is using that low vdimm
At what speed was it stable ?
btw can somebody verify the cpu-z screenshot ??
I only saw a bench @ 5,3 GHz, so not so amazing, cuz hipro5 made these clocks on his mach2 modded, too.
Quote:
Originally Posted by pc ice
Tis cos 90nm power consumptions scales out of control with frequency (ie heat output) cos of leaking transistors. Idle they can do very high speeds, put load on and the heat output at 6ghz+ will exceed 200 watts, power systems fold, the chip folds with it.
cpu-z screen checks out.
Regards
John.
cant leaky transistors be fixed?Quote:
Originally Posted by froudeg
They're inside the chip, and theres probably about 100million of them that you'd have to fix. Still up for it? :DQuote:
Originally Posted by Playful_Buffalo
something they could try is changing/upgrading the caps, that might help them, but not much.
Unfortunatly it took a while to create my account, to give you answers... ;)
CPU: 570J
MoBo: Abit AA8XE (not the Fatty), without a mod
Ram: OCZ PC2 5400
Cooling: LN
We tried additionaly the DFI Lanparty 875P-T, but didn´t get it over 300 Mhz FSB. The only BIOS that works with the 570J is a beta-version, but you could not adjust the NB-Voltage... :rolleyes:
We put our focus on max frequenzy, and didn´t test SuperPi at the Abit-system... :D
But more tests are planned!
More pics under: Forumbase
http://www.computerbase.de/forum/att...chmentid=19741
^^^^welcome to XS
i ment by intel :p in the fab processQuote:
Originally Posted by lalPOOO
The Coolly`s with DDR² Bord.
It´s real, it makes large fun and it continues!!
is that aluminum and brass?
6 GHz is plain MAD!
if only power is a limit maybe we should make new one ;). I'm sure that on this forum is enaugh specialists to make brand new mobo supplying :>Quote:
Originally Posted by enzoR
It´s copper and brass.
what an awesome result ! O_O makes me speechless ! congratz to the clockers !
Sweet looking containers there Otternase :)
New copper soil for the next test!!!
Amazing stuff u guys are doing there.. love how all the caps around the cpu froze..
u guys sure its sucking up close to 300watts at load. that seems a bit INSANE. boy that would put a load on any cooling system. ln2 or not.
At the rate intel is going if you oc one mhz u have to go to water or above.
very cute stuff guys ! :toast:
the next test can come :YIPPIE:
WAY TO GO!!! woot!
To give you an update:
We made on sunday a pre-test with a Fatal1ty, but not a final shot:
Keep up the good work
Quote:
Originally Posted by moddolicous
Changing caps wont do anything but waste money in this application. Caps dont provide main power and dont support voltage in a switchmode power supply. They ONLY filter the high frequency ripple, nothing more. You'd have to get Farads of capacitance to provide the current needed for a couple of milliseconds of voltage support when drawing this amount of amperage. And again, the caps have to be charged, they arent like a battery. THey become a load on the power supply after they discharge to the voltage of the supply.
If your power supply aint cutting it, Industrial power critical system SMPS's could be made to work with a computer pretty easily. The amount of current would only be limited by your bank account. I wouldnt trust anyone on these forums to do it based on what I read here everyday though. non-isolated High current electricity and computer nerds do not mix.
The motherboard could be modded to lower the parasitic resistances that are making the voltages fluctuate madly at these high currents. The toroidal output inductors for the SMPS's are not very large and high in resistance considering the current flowing through them. Winding larger (Size not value)custom made inductors would be very beneficial, but you'd have to get someone who knows what they are doing or else you'll be hurting ripple performance too much and risk a destroying a cpu. Winding several small strands of copper is much more effective than one large diameter solid core wire. I would go slightly less inductance to increase transient response of the supply and reduce the series impedance of the inductance.
Not sure, but the motherboard regulation circuitry probably has a built in current sense/limit on the Vcore line which would need to be removed/modified. At high currents, circuit board traces become a series resistor in line with the load. Something would need to be done about the path current takes in order to reduce resistance as much as possible. When dealing with small voltages and huge currents, 1 milliohm translates in to a very large voltage drop percentage wise. This is the reason for voltage sags on the Vcore line on any well designed motherboard.
More/Better filtering will be needed to compensate for larger ripple due to higher current draw and modified inductors.
This is the basic outline of what is needed to supply more current to the CPU. I'm not sure if anyone on these forums has the electronics knowledge to do these types of mods in a way worth doing... I hope I'm wrong. I'd like to see it happen.
Quote:
Originally Posted by stummerwinter
Sure looks nice! :toast:
But quite slow time for that huge numbers?
As written, just a pre-test without optimized timings... :D
We still waiting for additional components...;)
Now this is someone who absolutely knows what their talking about. Thank you this type of knowledge is not only beneficial, but is most refreshing when theory is based on fact rather then conjecture. best response I've read to a post in some time.Quote:
Originally Posted by SPL15
Hippro has a degree in that field...
Very nice clocks, grats
That pi is too slow. unpatched I can match that at 5.6Ghz maybe 5.5ish on the DFI or Asus. the dothan at 3.07Ghz pulls 25s
memory on 3:2?
SPL15: I fainted after i read that....
I forgot about Hippro... He should definately be able to pull those mods off.
I'm sure I could, but I've got enough contracts at the moment to keep me busy for the next 2 years. And plus my hourly rate is quite high because the companies I work with have very deep pockets.
Lets hope Hippro has some spare time on his hands.
@Fugger: Thanx ;)
Yes, unpatched, to be honest, we didn´t konw up to the weekend, that a patch exists... :rolleyes:
CPU:RAM = 1:1 (instable RAM at higher FSB, I think the Bios ist still not really developed)
I made a test yesterday at this system with FSB 220 / 3:4 ->
unpatched: 32 s
patched: 29 s :D
So, next time we hope to perform better!
good luck and uncanny setup
keep it going....
This Weekend :rolleyes:
cpu-z 1.27.1 pls ;)
i second that !Quote:
Originally Posted by saaya
that 4th picture looks like something out of et :Rofl:
cpu-z 1.27.1 is not working correctly.
We had some Problems while reading out the rigth frequency with 1.27.1, so we went back to 1.26, additional test for SuperPi are plannend within the next few days...;)
cpuz 1.27 doesnt work corectly? you mean it shows a lower cpu speed? ;) :D
i never heard that cpuz doesnt work correctly...
theres cpuz 1.27.2 out btw check if that fuixes the problem :)
whats the max superpi stable? theres a new superpi version with a checksum, thats also ok for verifying your oc if you cant get cpuz working for some reason...
what exactly is the problem with cpuz? it wont launch? shows wrong speeds? cant be verified/submit the result?
If the problem with version 1.27.x is the 14x multiplier, this is not a bug.
This article explains the change between 1.26 and 1.27 :
http://www.x86-secret.com/index.php?...=newsd&nid=846
If the problem is something else please let me know, version 1.27 still has lot of bugs, so please don't hesitate :-)
Franck
XIP WITH ONLY 24 POSTS !!!! GOOD GOD !!!!Quote:
cpuz
X.I.P.
Join Date: Jul 2004
Location: New Jersey, USA
Posts: 24
Unless you're the creator of CPUZ, that lovely program which I stare at everytime someone posts a benchie, its fine then :D
Perkam
he is :)
unbelievable its all done by one man, no? :D