Printable View
Yep thats pretty much how you can tell if your board is going to run blend at a given fsb. If my board doesnt pass memtest(dos) test 5 without errors then it wont pass blend. If it restarts itself or doesnt finish running then your having serious nb cooling issues.Quote:
Originally Posted by shimmishim
Quote:
Originally Posted by Navig
Hey navig, try 1.4-1.5v on the cpu, then 1.75 on the NB.
you usually wont be able to touch 3.6ghz on stock volts like you are trying to do.
Thanks for the advice, I'm making another run here. I'll try the higher voltages again. However, I actually don't think I'll be changing out my northbridge. I happend to pick up a laser thermometer. Under orthos blend, the base of my custom nb heatsink peaks at about 29c, so that seems alright to me.
navig
Just fyi, I am running the new version of P95 to test stability. I know that my OC is far from spectacular, but I just want to share that 400FSB is a-ok with this board (actually, even with my 9/15/2006 official bios).
Anyway, my BIOS is the latest one posted by NicoOCZ. It didn't break my board, and max FSB is the same. I don't know about Orthos testing, I never really liked it so I haven't tested it and I have no plan of using it.
Bigger Image: http://sleektech.nl/~dm/temp/abit/abit-400fsb_lg.jpg
http://sleektech.nl/~dm/temp/abit/abit-400fsb_sm.jpg
It's just one-hour testing (but it's still running till this posting). It's running Blend, and for some reason, it says 399MHz FSB when I did set 400FSB in the BIOS. In any case, vMCH is set to 1.6v, the rest of the voltage and tem is in the screenshot so just check it out.
I can't push this memory beyond DDR-830MHz because that'll require me to go 2.4v and higher, right now, my DDR-533 is doing good at 2.2v.
Here's a CPU-Z: http://valid.x86-secret.com/show_oc?id=149461
http://valid.x86-secret.com/cache/149461.png
CPU : Intel Core 2 Duo E6400 (*1)
CPU Arch : 2 Cores - 2 Threads
CPU PSN : Intel(R) Core(TM)2 CPU 6400 @ 2.13GHz (ES)
CPU EXT : MMX SSE SSE2 SSE3 SSSE3 EM64T
CPU Cache : L1 : 2 x 32/2 x 32 KB - L2 : 2048 KB
Core : Conroe (65 nm) / Revision : B1
CPUID : 6.F.5 / Extended : 6.F
Freq : 2394.7 MHz (399.12 * 6)
----------------------------------
MB Brand : http://www.abit.com.tw/
MB Model : AW9D-MAX (Intel i975-ICH7)
NB : Intel i975X rev C0
SB : Intel 82801GB (ICH7) rev 01
----------------------------------
RAM Size : 1024 MB
RAM Freq : 399.1 MHz
RAM Type : DDR2-SDRAM Dual Channel
RAM Ratio : 1:1
RAM Timings : 4-4-4-12
----------------------------------
Slot 1 : 512MB (PC2-4300)
Slot 1 Manufacturer : Micron Technology
Slot 2 : 512MB (PC2-4300)
Slot 2 Manufacturer : Micron Technology
hey, i had these supertalent as well but had trouble w/ them on this board over a certain speed...memtest errors, couldnt run dual pi, and also w/ these ram modules and the board i cant put it at any voltage above 2.1, if i do it blue screens and doesnt load into windows. With 2.1 i got it to 3.4..even 3.5, but not memtest stable or dual pi stable. I'm thinking its this ram on this board or somethin...i rma'd the super talent and i just got the Team Group 800 cas4...will install them this weekend. Look at the Extreme Bandwith forum on here and you'll see the post about these modules.Quote:
Originally Posted by shimmishim
I have the posted bios also. The flash went without a problem. I just need better cooling on the cpu. The stock Intel hsf is crap. I am going to have to get a waterblock for it.
tested everest read and my board is compleat stable, i can go higher mvch 1,75-vfsb 1,35v - memv 2,35v 4-4-4-8 and all @ air
http://img102.imageshack.us/img102/7...readhb1.th.jpg
Well, just ran ORTHOS Blend for 12hours at 267FSB the whole night. No errors whatsoever, everything at stock volt.
For some reasons, I suspect its the TRC that could be causing the ORTHOS Blend stability. Currently, its at TRC 20.
Btw, BIOS 13 Final.
Well I tried for 9*379, but I couldn't even pass dual sp 32m, despite a core of 1.4375, MCH 1.72V, DDR 2.3V. Did manage 3dmark05 and 06.
I don't know whether I'm lucky or not, but on this abit(I kicked my P5B Deluxe for it was lacking any control over the fans) I can run without any problems 9x400 @ 1.65 MCH and 1.375V CPU and 2.0V memory(E6600, cellshock). It is as stable as it gets, but to push it higher(420x9) I have to push the MCH @ 1.75V, and I'm not confortable with that. At the same time, I'm not in for peanuts, so 405 vs 400 in my eyes is the same.
Look, VMCH on this board droops alot, so grab your multimeter and measure it.Quote:
Originally Posted by roibm
If you need a measuring spot on the front of the board, I have one...
2.0v bios with newest Xeon = 1.87v MCH measured, 1.91v with my other cpus.
1.75v bios is prolly 1.6xv in real life,
Can you post measuring spots? TnxQuote:
Originally Posted by fhpchris
Quote:
Originally Posted by <><
wow, it was the memory...just popped in the Team Group modules, set them at 2.25v, 4-3-3-5-40 timings...booted right up, dual pi stable....new super talent = no good!
Okay, found something interesting today.
First, my OC setup:
400FSB (399MHz actual)
5:4 (RAM at 500MHz)
1.75vMCH
BIOS 13 Final
5-5-5-15-42
OCZ PC2-8000 Ti Alpha VX2 @ 2.55V (blah!)
So, I ran Memtest 1.65 (bootable) and it gave me about 1000 errors in test #5 (first loop). This was with the stock NB heatpipe cooler. Although it was only 1.75vMCH, the NB heatsinks got too hot to touch even when I had a 120mm fan blowing at it. Damn...
Then, found my Thermalright HR-05 and mounted this badboy onto the NB. Tried running Memtest 1.65 (with above setup) and finally no more errors in test #5. Yey!!!
Currently doing 500MHz memory at 4-4-4-12-42, and still no errors in test #5 yet. Passed 5 loops in test #5, no errors at 4-4-4!
The HR-05 is currently passive but I have a 120mm fan blowing towards it. So it gets a little warm at 1.75vMCH. I guess BIOS 13 Final is overvolting the vMCH a bit... Will update with Windowz testing soon. Anyways, changing the NB heatsink does help improve overclocking alot. Got me DDR1000 easily!
^^ running mems @ 4:5 is not so hard like you said. 1:1 is much more so. If we can just figure out how to get around the difference. It could be the Trc or a balance in voltage/cooling. I tend to agree it has something to do with timing though. edit: I always run my vdimm and vcore high as I have good cooling and only use it for short periods at a time.
yeah it's been said about 100 times that if you are serious about clocking this board it absolutly needs better nb cooling.
Another problem with the board is how much further it droops vmch the higher you go with it. I need 2.0 to run blend stable at 423fsb but its really only about 1.86-1.89.Quote:
Originally Posted by pcmoddingmy
agreed Mrdeeds...
that's why this board needs the vmch booster mod.
i've done it and it makes the droop only .03 volts.
so 1.85 = 1.82 actual. i've measured it before too.
also... to the person who ran blend at 267... it's not low fsb's where the problems occur.. it's above 400 where it becomes an issue.
the highest i've been able to do blend is 405 w/o errors for 20 minutes... then i stopped and got a little greedy...
the super talents work great... like i said, it's the board that's the problem...
i felt like i wasted money buying another set of memory just to confirm that the problem was the board and not the memory.
i'm glad i'm selling this board. the p5b deluxe will hold me over until someone can make a good/stable overclocking chipset (possibly rd600? new dfi 680i?)...
also... i really wonder why abit doesn't want to support this board very much... 1 bios update in 3 months seems ridiculous considering Asus released like 3 or so within 3 months for their boards...
also, this board (despite having all the bells and whistles) definitely does not deserve the "max" branding... you'd think they'd provide more support for overclockers...
try the new p95, i bet its stable for you.Quote:
Originally Posted by shimmishim
Don't use orthos, several people have found that orthos isnt stable where everything else it.
thanks for the suggestion.Quote:
Originally Posted by Revv23
i actually gave the new multi-threaded version of prime95 a try and my blend errors at exactly 6 minutes...
it always seem to get stuck at the 3rd loop of the 1024fft length.
I asked at another forum about orthos as a stability tool, Johnny Lee gave us his verdict whch means there are no real hardware conflicts known on the newer platforms if you like Orthos as a tool. Other than that personal stability remains a preference..
An interesting read none the less.
http://www.dfi-street.com/forum/show...0&postcount=21
All I have found on the newer boards is that some run Orthos better than others, especially at higher fsbs over 420. For the most part I take that as some mark of having things right.
Quote:
Originally Posted by shimmishim
Fair enough, at least you tried right? :)
Merry christmas. :) :toast:
do you have a link to this new version of prime?Quote:
Originally Posted by shimmishim
eva's thread on XS
http://www.xtremesystems.org/forums/...d.php?t=126820
Yep, Prime95, always been the best stability test software even before the new version came. Tried and tested!