Thanks guys.

I was just going by what I thought Splave said....but what he actually said was "1.6vtt (board has bad droop, this will equal low 1.5's)" .....it just showed 1.65 in the s/c.

I'll definitely go lower today on the VTT(for starters anyways).

I have eist enabled so I can have Turbo enabled for max clock speed.

Just disable C-STATES then, though I never heard of needing EIST enabled for turbo to work on X58.

^ Nor have I...possibly related to that specific mb?

If I disable EIST, turbo disappears from the menu.

check it with a dmm ;) you will be shocked

Im tellin ya guys you need eist lol I just benched it for 2 hours to get him some settings :) xpower is almost as ass backwards as this deluxe p8p board...
What bios you on NY

We believe you splave, just never heard of it before SB. :blink: Last MSI board I ran was the X58 platinum 222BCLK wall but was very solid up to that and disabling turbo opened up the 21X multi for all 4 cores in BIOS IIRC.

Which one on the board is VTT....is it QPI? If so I get 1.13v at stock settings. The others are CPU-.95v, DDR-1.49v, IOH-1.09v, and ICH-1.48v

I have BIOS V1.4B8 according to cpuz.

WOW.....assuming QPI is VTT at this point...I set it to 1.5v and get 1.28 on the meter. I'm running wprime1024 now and the cpu is not getting too nhot at all like last night. Max temp is only 43C today and it hit in the 90s quickly last night with 1.65VTT.

Wprime is still going, 30% now 44C max. Guess 1.65 isn't necessary at only 220 bclk/4620 cpu.

EDIT: that was only 1 thread..oops, with 4 threads it hit 80C pretty quickly and I stopped. What is too hot.

ICH should also be 1.1v like IOH. More than likely that reading is misnamed and it is the PCIE voltage it is mesauring.

Yes VTT = QPI or QPI DRAM depending on what company produces it.

Good to know thank you.:thumbsup:

I just ran spi32m @ 230 bclk with VTT set to 1.5v and reading 1.28 @ idle and PCIE @ only 110MHz.

It froze up when I increased it another 2MHz. Not sure what was the cause yet.

Memory was @ 900 something with 8-8-8, going to try 8-9-8, but probably need more VTT or PCIE to get higher than 230 bclk.

EDIT: looks like I'll need to get some LN2 for the lengthy benches, or learn DICE.

Sorry everyone....it turns out when I was saying I was using 1.65VTT, what I meant was CPUPLL. I was just confused. I benched spi with 230 bclk and 1.35(1.28dmm)VTT.

Now I'm going to try raising the VTT a little and see if I can get past 231 bclk.:thumbsup:

And check this out.......I made it so I can easily check voltages right from my seat.
Left to right in back is cpu, qpi, & dram, and in front is ioh, ich, and gnd. :cool3:

stop drinking dave!!! lol

How come when I set it to 1 core and change nothing else it BSOD, but with 4 cores it boots up np?

because it is 22x....need more vcore...

turbo mode: +2x on


...what is your bclk?

Booting @ 220 till I set it to one core.

i had that happening to me too!
and when i have all 6 on it works fine

If you just want to use one core why not just do it in Task Manager? What is it you are doing that would make you want to boot with one core?

What I do with my 4-way is boot in @ 215 around 4.5ghz I think, then use software and goto 4.9ghz+. Saves me from worrying about weather or not I can boot into windows or not. Although it's a evga board, with my MSI p55 I have to use the bios for overclocking. If you could get away with software Ocing I'd do that.

I guess that'd work. Thanks.

EDIT: 215 didn't work but 210 did. -2C with only 1 going.

Ohhh wow...I get a 22 multi now and only had a 21 before.......sweet and it's staying @ 22......ahh but it crashed at only like 220 bclk....not stable for some reason.

UPDATE: So for those that are wondering, I managed to get it to 5132MHz-244bclk with ~1.55VTT. Haven't tried to press on yet, but temps are safe (~10C) without being loaded.

22X multi + it just sucks on single core mode. BCLK wall goes WAY down.

Very nice on the new CPUZ... 1.55VTT You mean Vcore?


22X multi... it never worked for me... but not trying stuff isn't exactly the OCA way :)

benching with the soundcard? lol

I was wondering what the heck that was doing there :taunt:

Vcore and VTT.:thumbsup:

I don't really have a dedicated daily driver......well on 2nd thought my laptop , but I can't put that soundcard in my laptop. ;)

nice chip! :D faster then mine on single stage

There's more where that came from.:taunt:
Might try a little more vcore or something for 1 more run before I tear it down and wait for new HD and the rest of my triple channel kit.

Also need to check under the block for moisture....the chiller has been on there a few days now and it was my first time sealing one of these in. All I used was eraser.

EDIT: here's what it looked like. Just the slightest bit of dampness on the towels right around the chip. Perfectly dry and sealed in tight IMHO. :cool3:

Check out the thermal paste though, not sure how it got like that. I did use more than normal for some reason, but it looks weird. Has to be caused my moisture. I've never had that happen before, and I know it was squished down tight to the cpu....the surface of this chip is nowhere near as smooth as an AMD.

afterthought: I don't think it was seated good even though the temps were pretty good. Under load @ 4.8Ghz they were outrageous. There cannot be any insulation between the head of the block and the hold-down plate. NP.

Oh, ok. That makes sense :thumbsup:

So I just tried running 32m @ 5100MHz(bclk 243)........did fine @ 5000MHz w/ only 1.506 vcore and 1.547 qpi.
1.537 vcore,
1.555 qpi,
1.685 ddr,
1.197 ioh,
1.540 ich
memory @ 971MHz 7-9-7-23-78
qpi link @ 2914
NB @ 4372.

It crashed BSOD ~ halfway through 32m. I then proceeded to lower the NB a notch to 3886 and it crashed BSOD part way through copy waza.

Any suggestions?

Thanks.:)

raise trfc a couple notches?
what was the bsod code

No code it was just a quick BSOD then restart with failed OC messages.

The memory is fine in 32m up to 1050MHz @ those timings.

EDIT: ohh yeah I need to set it so it does not auto-restart upon BSOD.

I live for double posting.......

Anyways I got it to run pifast @
cpu=5100MHz
uncore/nb=4370
qpi link=2900
ddr=971MHz/7-9-7-23-78

vcore=1.53v
uncore/qpi=1.33v(dunno why I had it so high before)
ddr=1.69v
ioh=1.197v
ich=1.54v

(All these voltages may still be higher than necessary and causing insability)

Haven't tried spi yet with less uncore/qpi voltage. I should be able to run it at least as high because I'll only be using 1 core for that.

I'll report back later.

try ioh/ich 1.25-1.35v both the same

Well guys, maybe 104 points is all I can do with the cooling I have available. I might get like 10-20 more points with better cooling.

When I'm single core benching it's hitting ~15C and with wprime it's hitting like 50-60C. I need to get a better water-block for the chiller for one thing.

So I seem to be banging my head on the wall now. I just don't know what to increase and it seems like a million possibilities.......along with the possibility that it just may not go any further without being subzero.

I'm sure I'll keep trying, but what do you guys think?


http://hwbot.org/community/submissio...920_5184.9_mhz

http://hwbot.org/community/submissio..._920_16.56_sec

http://hwbot.org/community/submissio...920_7sec_969ms

http://hwbot.org/community/submissio...in_19sec_516ms

http://hwbot.org/community/submissio..._920_5sec_54ms

http://hwbot.org/community/submissio...in_40sec_524ms

i say u did an awsome job to be just on a chiller :thumbsup:

Thanks DOM.:)

I guess I should get some 3d benches installed tomorrow, I have some 3d studying to do though.

I have 3 unmodded gpus I can bench on chilled water.........4890, 4850, and 8800GTS 640. Should be able to scrounge up a few points there, even with a 920........I hope.

Yeah, no shit. Keep it up there NY. :thumbsup:

Awesome 5ghz club member on chiller, nice work bud :)

Awesome results! The 920 is a real bointiful cpu ;)

Thanks brothers!! I appreciate the kind words, and I'll do my best to keep rolling in the boints.

When I asked "what do you guys think", I kinda meant more what do you think about the likelihood of my chip being near or at it maximum potential using only my chiller.

Do you think I can get any more out of it without some LN2 or DICE action?

on ln2 you should get more mhz but wouldnt know how much more.... depends on the chip

ICH should not need more than 1.2-1.25 IOH though you can go higher with. It helps with stability at high BCLK.

IOH 1.1 @ 1.35-1.45
IOH 1.5 @ 1.6-1.65 (helps with 3D) this is also called IOH IO IOH/ICH IO or PCIE voltage.