youngpro

Evening all,

Today Cera and I finally got a chance to bench some DICE with my qx6700, I've been waiting for this for ages. Unfortunatly my commando is still out of action, so ill be benching on the p5k deluxe. This is good and bad, not as good voltage options, but the p5k allows might higher fsb for the qx6700 and this can mean better PI times.

Bench Setup:
QX6700 L644F040
ASUS P5K Deluxe + vcore mod
Cellshock PC8000 C4
evga FX5500
Kayl Dice/LN2 tube
Vapochill LS (stock) on Northbridge
Thermaltake Extreme Spirit II on Southbridge
Thermalright HR-09U on Mosfets
OCZ Gamexstream 700W

Temps were good from the start, bottom of the dice tube was reading -66 and preety much stayed there all day, when volts got as high as 1.75 it lowered slightly to -64 and under some wprime runs it went down to -58.

The board handled great, didn't really hit any fsb walls throughout, held the volts well without hardly any droop.

I like to cut to the chase so here are some results and pictures

Max clocks on DICE were 4768MHZ. This was done with vmch 1.7, vcore 1.7, vsb 1.2, vfsb 1.6. Temps were still about -64 at the bottom of the dice tube. Unfortunatly I've got one of those broken temp sensors that never show below 9 degrees in any monitoring application.

http://valid.x86-secret.com/show_oc?id=213271

I was also quite happy to get a 10.671s PI time, this was done at, 4740mhz, 474 * 10, memory at a 4:5 divider at 592.5mhz running 4-4-4-8, subtimings 20-6-10-10-3-7-4. A few services shut down, pi running in real time, system cache optimization, I obviously didn't do enough because my efficiency is only 50580



I've got some more results to post shortly, an excellent wprime run and a quite exception 6400 3DMark06 CPU score, I'm not sure how I snagged that one. Most of my max scores about 4550mhz were 5369, after some memory timing optimization one of my CPU runs scored the mid 6400 score, but when I get some air cooling back on that box I'll retrieve and post that photo.

eva2000

wow congrats man 4.78ghz for QX6700!

Cera who ???

So you could 3dmark06 @4.78ghz ? Managed to pull of any Super Pi 32M ??

Love the accompanying photos

kayl

Not to bad hey , cant believe ya managed to convince the girlfriend to bench
Did she have candles for the occasion
Love the vapour LS on the Nb as well

youngpro

yeah 4.78 isnt bad, should scale nicely under ln2 i would have thought,

nah 3dmark maximums were about 4.6ghz, we didnt get time for pi 32mb, but at the 4740 i was running pi 1mb it was stable enough to do some 32mb runs, but next time i guess

haha yeah i know kayl, well she thought she would be bored at home and ended up enjoying it, she was increasing setfsb for me in windows and testing stability with pi unfortunatly no candles

in the end i'm not sure having the vapols on the nb did alot, my memory timings at high clocks seemed better, and i was able to squeeze another 3mhz out of memory at 4-4-4-8 but that could of just been ambients or other conditions, was fun though

dinos22

nice going man wow

moloko

Heheh, gotta love playing with DIce!

Nice work mate

Is there going to be a chapter 2?

I love the vapo on the NB, it looks ace

I've been playing with some dice for the past two weekends (10kg's first weekend, 20kg's the next). Some neat results (E6700 @ 4.65 w/ 1.62v) though I'll post more complete info in due time

I'm really surprised that you were able to hold such low temps with the quad on those voltages. I only got time to test 1 of the 2 quads I had this weekend and it just ate the DIce whilst running 1.7-1.8v on the P5K Premium. Quite literally the DIce was just sinking into the tube. Not quite as dramatically as that sounds but any one who has benched with DIce, imagine being able to visually see it sinking rather then the usual slow sublime to CO2.

youngpro, I may be in contact with you soon also mate, I'm cooking up an interesting benching concept


I didn't have a vapo for the nb, but I damn well tried with a metal shot glass and some dice! Too bad it failed like a fat kid trying to resist a donut.



Here's the dice pot at idle:



But hit load and it'd just go nuts (-44 to -36, higher at times):



Then packed again whilst under load:



Only to have it go and sublime so quickly:



Unfortunately I didn't have any acetone so instead I used methylated spirits. Much the same effect and here it is after finishing off the last bits of dice for the session:



It's a crap photo but I wasn't using too little or too much, in my opinion at least.

I assume you used some sort of alcohol also youngpro - what did you use and how much?

DIce is fun, hurry up Aus and get into it

youngpro

well moloko that vapo on the nb looks good, but im my case it didnt do alot, i think it would make more impact on the 6*50 chips where the motherboard is limiting the fsb and not the chip itself... i think with all the 1067mhz chips your just as well off using water cooling on the northbridge as you are running a dice tube or phase change..

congrats on the 4.65 on the 6700, thats a good result!

i know what you mean about the quad eating dice at high temps, at one stage i looked away for a minute and then i saw 25 on the temp monitoring, i thought sh*t -25 is getting a bit high, when i relooked it was reading +25, in about 5 minutes it at all the dice in my tube!

were you running a volt modded p5k deluxe, i saw you mention you did up to 1.8v, i didnt go above 1.75 as it really was just wasnt improving oc..

feel free to message me about any ideas, always up to try something new..

yeah well mate hopefully when i get back from my holiday in about 3 weeks ill be benching ln2, ill start with a celeron or something to get the hang of it, then itll be qx6700 and 8800 (kayls cooking something up in terms of sli 8800 tubes)

dinos22

moloko that doesn't sound right about temps hmmmm