Printable View
Kentsfield K6700es B1
ASUS P5W64 WS Pro , all stock, no mod.
Team DDR2_1000 1G*2 Victor ED
Zippy PSL-6701p
Raptor 150G
FX5700 128m PCI
setting:
vcore=1.45v
vmch=1.8v
vfsb=1.2v
vdimm=2.4v
multi=9x
365*9=3285mhz pi_1m=13.875sec , I really cannot believe that even it's work out by my own hands.:eek:
I'll tell u guys what happend tomorrow :D
Hmmm, i think the record is like 9seconds ? lol but never the less that is a great result, my X6800 does the 1M one second longer than your socre. but your FSB is much higher than mine, i just bumped the Multiplier to 13.
Thats a pretty nice result for such low clocks.
so jack it up with LN2 and let's see this at 5GHz and 8.x sec run :)
so how's you do it :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D :D PM ok too hahahh
I think you haven't notice how fast it is, ONLY 3289mhz pi_1m = 13.875sec :p:Quote:
Originally Posted by Artmic
13.875sec need almost 3.65G to achieve in common. :p:
you will know it by tomorrow :DQuote:
Originally Posted by dinos22
Almost seems like PI is trying to use more than one core?! :confused:Quote:
Originally Posted by VictorWang
That is an excellent time for such low clocks indeed!
Tell us How :P
you can trust me it's a real clean run with optimized ram_timings and OS, but not by PS :cool:Quote:
Originally Posted by DeToNaToR.cl
well for me it is "tomorrow" officially heheheh 1.22am in Sydney as i type sooooooooo i'm ready to learn from the masterQuote:
Originally Posted by VictorWang
please don't tell me it's a bug of some sort
do another one with different RAM and CPU speed.....i want to see if this is repeatable
you able to run 32M with maxmem at 600?.....i was testing ~250 maxmem recently and it would just freeze SPI in 32M runs....hmm
how are 32M runs with this tweak or whatever you figured out
i think it's a bug...
SuperPI isn't an multi thread soft...
http://wowgame.free.fr/Conroe%20E660...013.937s-2.JPG
Maybe that "Multiplexing" thing?
Think it's a bug too. The 2x4Mo cache is not enough to explain the difference of 400Mhz for the same score ... or you've found the ultimate optimization :clap: . I want to learn more about it !! :banana:
ok man , you beat me :DQuote:
Originally Posted by dinos22
as I tried so many time, it's a BUG when K6700 meet P5W64
You can see the following 2 pics , then you will know what's happed :D
maybe that's why ASUS 975x+6800 cannot run at 400fsb+ on normal Air cooling :p:
setting:
vcore=1.45v
vmch=1.8v
vfsb=1.2v
vdimm=2.4v
multi=10x
1. K6700 @ 365*10 1095 4-4-3-1 pi_1m
http://www.xtremesystems.org/forums/...1&d=1157989167
2. X6800 @ 365*10 1095 4-4-3-1 pi_1m
http://www.xtremesystems.org/forums/...2&d=1157989233
multi drops to 9x instead of 6x hehehehehe
no , I set 9x in bios, but... sometime it's running slower than normal, but sometime is running extremely fast :DQuote:
Originally Posted by dinos22
So I gave you the "FAST" result and compare with performance of deafult 10x :p:
oh the other way round ROFLQuote:
Originally Posted by VictorWang
nice one anyway you were :banana::banana::banana::banana:ting your pants i'm sure until you figured it out hahahahaha :toast: :toast: :toast: :toast: :toast: :toast: :toast: :toast: :toast: :toast: :toast:
I've read there's some errors with application for multithreading on Kentsfield.
Is it right? And GOOD overclock for air.. :woot:
It is because, there are two dies on a package. the changing multplyer is only effected on a die, the other die is still running at 10x.Quote:
Originally Posted by VictorWang
You can ask manufacture to correct the BIOS.
I got three results at continuouse three runs.
Bios still can not handle the cahanging of the multiplyer on die No.2(on this case CPU2,3, still run at 9x266)
haha, Asus' great BUG, and great you found the reason :toast:Quote:
Originally Posted by tanghiro
weird bug heheQuote:
Originally Posted by VictorWang
VictroWang, are there any known wmods for this wonderfull mobo?
Can you publish or send some high-res pics to sb like shamino to figure out some mods?