Sub 4 GHz Pi Thread

[Vapor]

We got 12.515 a few different ways, this was our fastest 19th iteration run (11.968....the others were 11.984 or 11.983 IIRC), but still couldn't get 12.000

Those were practically the only two timings we didn't try loosening...wow, lol.

I have it all right here....and this is tempting me to give it another go

I think I'll let the board and RAM rest until next weekend though

...unless Gautam wants to give it another go when he comes by to pick up his stuff

[MidnightMadness]

12.516s - 3987MHz



12.500s - 4005MHz



Mobo: Abit AB9 QuadGT
Mem: Crucial 667 10AE

[Dekara]

@Madness

Really great spi time, your win is fully tweaked, congrats!

[MidnightMadness]

Thanks I try as I can.

[staxm]

mobo Abit QuadGT

[Gautam]

Great times MidnightMadness and staxm!

[killert]

Im going to be laughed out of this thread

[G H Z]

Took the DFI 965 board for a spin

[Gautam]

lol come on bro, just .001 more.

[JMKS]

Some even numbers with Allendale @ 4GHz .
Not the best possible with this frequency (maybe 50ms slower than it can be), but the zeros are nice .

[Gautam]

If enough people come in with 2MB cache times, I could add a second category...

That looks good, only about 1.5 secs slower than 4MB.

[G H Z]

lol come on bro, just .001 more.

For as many times as I had to click 'begin calculation' to get this result I don't care about 1/100th of a second!!! What I really wanna figure out is 4:5 on this board.

[Solarfall]

just noticed this thread.. mmm i think i'm going to do some superpi tweaking today

[yotomeczek]

Abit QuadGT, WC (water + 8 degree)

[Gautam]

Nice time Yotomeczek.

[yotomeczek]

Nice time Yotomeczek.

THx
I have problems with 16&32M I loose abot 5s in 16M - don't know why (I did 6m14,5 @ 7x500 and now I can only 6m16s ... but Madness can do even 6m12,5 )

[Gautam]

At the risk of looking like a jerk, I'm afraid I'm going to have to take off before's time after all. Sorry Xavier. I tried 4003.6MHz and it is just too fast compared to 4000MHz. Not a fair fight, so I have to remove it.

The reason I'm going to look like a jerk is because I've got a new time too.

[hausner]

nice result Gautam !

[Misko]

Really nice one Gautam

[dinos22]

looks like i'm on par with 3oh6 with 444MHz 3-4-3-4 timings as i've only got a 9xmulti

a touch quicker though and i know i can go down to ~ 12.55 with maxmem=500 >> maybe more with a smaller value

[3oh6]

looks like i'm on par with 3oh6 with 444MHz 3-4-3-4 timings as i've only got a 9xmulti

a touch quicker though and i know i can go down to ~ 12.55 with maxmem=500 >> maybe more with a smaller value

nicely done...that is a good pinch better but no memset in the screenshot .

i was hoping to get some time to run this again in the near future 500*8 (quadGT). see what 1000 4-4-3-4 does at 4 GHz on that board. i should get to it in the next few weeks i hope.

i would think that to be faster than 444*9 at 3-4-3-4 that we can run...yes? no? my P5B-Dlx will just not run any faster than 475FSB on the tight strap after modding and cooling with water. the quadGT did 500FSB out of the box on stock cooling.

enough rambling...nice work though, sorry for the off-topic

[dinos22]

yeah QuadGT will beat it easy..............it seems that motherboards has killer SuperPI performance

sorry my RAM timings were not all that special to be honest......i haven't really tweaked the run yet >>> i've only discovered last night at 2am that my CPU can do 32M SuperPI at 4GHz.....kinda exciting hahahha

[andre X_X]

hmm... 1:1 with 9x multi and 1066 strap wasn't bad

[Vapor]

Damn....I wouldn't expect 9x444 1:1 4-4-4 to be that fast at all, must be the low available memory....how do you get it that low?

[andre X_X]

dont know why it would be that low... this is the setting/tweaks for 32m (except wazza) that i used to race with gautam in low clock challenge, but usuallly i got 4xxx on available memory (with wazza). i mean, i use the same tweaks/setting with 32m run, excetp the wazza...

this 12.531s was the best i could get with 9x multi, no matter 1:1 or 4:5 the result was the same, except 1:1 @ 1066 strap have more chance to get 12.531s. i'll try 8x500 1:1 @ 1066 later on.