QX9650 on a EVGA 680i SLI

[ncc-1709]

http://valid.x86-secret.com/show_oc.php?id=280447

that should shut you non belevers up before it starts.
ALL 4 cores
http://valid.x86-secret.com/show_oc.php?id=280249


now. im looking for some advice on ram settings to stabilise my ram somemore before I go waisting $150 worth of dice trying to find out.

My ram is REAPER 9200.
i can run at 1260 ish on 5-4-4-10-T2-14-35-5-8-5-7.8u-5-3-2 on 2.3v.

it dont like much more ovltage, and concidering this is my 4th set, i dont really wanna give it any more.

now i would like to know which of the subtimings i should alter to improve my Pi time, then any other benchmark.

this is my best aqm3 so far, non Vmoded GTX's SLI air cooled.

[iadstudio]

according to cpu-z, ur only running 2 cores.

[pumbertot]

is the high multi and low FSB a consequence of Yorkfield and 680i or your choice?

and what BIOS is that you are running?

[cowie]

very good news...
two cores at 4.5ish on this board would be fine by me .
.

[iadstudio]

Everyone knew dual core would work fine on this board. It was Yorkfield support. With those settings ur essentially running an E series chip, not Quad. Nothing new there. I want to see 4 cores running!

[ncc-1709]

sorry, that was warm water testing.
http://valid.x86-secret.com/show_oc.php?id=280249
http://service.futuremark.com/compare?3dm06=4335832

the FSB is a bios problem (p31) as i can take it up to 400 with clock gen. but its limited to 320 in the bios. with 333 being the max it will post at.

and for AQM3 i only bench with 2 cores to save on my temps. trying to keep the temps down with all 4 cores is a pain in the ass big time

[iadstudio]

Nice. now that my friend makes me sick that I got rid of my 680i.

good job

[disruptfam]

can u confirm all 4 cores work on evga 680i

[Praz]

$1000.00+ quad-core processor running on two cores with mid 300FSB capability. Would hardly call that a success in reference to motherboard compatibility. nVidia would be in a lot better position if more of their users were so easily pleased.

[ncc-1709]

you not read the rest of the thread???

now, any one gonna give some tips?

[CrayS-per4m]

his CPU-validation shows 4 cores so whats the problem?

[Praz]

No problem. If a person is happy running that type of processor on a board that is only capable of sub-par FSB then we are all happy. But there is no need to make it seem as it's a breakthrough. That type of configuration has been able to be done since the processor was released.

[cowie]

well pleasently surprized is better to say than happy.

[dinos22]

great what bios is that on

[KTE]

Run 1200 4-4-4-4 PL5 if you can. Bench it, i.e. EVEREST, 1M, 32M and so forth.

Depends what volts you want to feed it. 2.4v might get you 650 4-4-4-4 PL5 benchable if you're lucky.

[dinos22]

what you want is 4-3-3-x on 680i

that is key to good RAM timings but forget about running superpi on 680i it's just SHYTE

[saaya]

yeah but hes on his 4th set of mem already and doesnt want to kill another one just to get 1% higher scores i think

[dinos22]

i always talk about benching rather than 24/7
since he mentioned superpi that means all bets are off hahahah

[ncc-1709]

well, thats PL5?
best i can manage on CL4 is 1100 at 4-4
i can squeze it down to 5-4-3-8 at 1220, but its slower in most 3D benches.

and for SPI being crap, well my 32mill is about 5 mins off the mark, so thats one of the reasons for the question.

[mdzcpa]

I feel compelled to jump in here and clarify that EVGA has made it clear that the quad yorkies are NOT supported by their 680i mobo. I don't want the uninformed reader to beleive for 1 second that the 680i is a workable platform for the yorkfield based on this thread.

By my own testing, and the testing of many many others, the support for a yorkfield on the 680i reference boards is non existent. I was able to get 2 cores working with a qx9650 on my evga 680i months ago....but so what. That's a dismal failure at best.

Furthermore there will be no magic bios to fix the issue. I don't mean to rain on the parade of the thread originator, but the thread is misleading at best. The 680i board needs physical modification in order to support the yorkfields.

Carry on.

[ncc-1709]

no physical alterations to make it work, no bios mods.
i tried it with the p29 bios first, but it wouldent boot(but would post)
so i flashed to P31 and foun that it wouldent boot again, so i lowered the fsb because i know that the p31 bios has a crap FSB limit, down to 1066 and it booted stright up, so i just worked from there and found the limit.
what evidence do you want as proof? camcorder footage?

[tombman]

no physical alterations to make it work, no bios mods.
i tried it with the p29 bios first, but it wouldent boot(but would post)
so i flashed to P31 and foun that it wouldent boot again, so i lowered the fsb because i know that the p31 bios has a crap FSB limit, down to 1066 and it booted stright up, so i just worked from there and found the limit.
what evidence do you want as proof? camcorder footage?

So you made it to the bios with the QX9650 installed and lowered the fsb down to 266 and only THEN was it stable?

Btw, can you prove 100% stability by posting a Prime95 4-thread (small fft) screenshot (running at least 30 minutes) with all 4 cores enabled?

PLEASE!

[ncc-1709]

i can prove by doing it, but i dont have anything like prime anywhere near this system. only F@H SMP client.

[dinos22]

i can prove by doing it, but i dont have anything like prime anywhere near this system. only F@H SMP client.

that's very hard on the system

[fragmasterMax]

no physical alterations to make it work, no bios mods.
i tried it with the p29 bios first, but it wouldent boot(but would post)
so i flashed to P31 and foun that it wouldent boot again, so i lowered the fsb because i know that the p31 bios has a crap FSB limit, down to 1066 and it booted stright up, so i just worked from there and found the limit.
what evidence do you want as proof? camcorder footage?

I hope this is possible with my 975xbx2, and the 965p's