im doing fine with the final bios. 455 fsb at 3.6 ghz, with a 8 multi(e6600). still messing around with it.
Printable View
im doing fine with the final bios. 455 fsb at 3.6 ghz, with a 8 multi(e6600). still messing around with it.
Yup, I've got that new prime95 too, but I still feel running ORTHOS blend makes my life a little easier.Quote:
Originally Posted by coop
Anyways, ORTHOS blend errored out at over 2 hours after I've switched TRFC from 35 to 42 at 401FSB.
I flashed from windows initially like i've done 20 other times before. Cleared cmos restarted and could no longer boot into windows. It would crash right after the windows screen stock with no oc. I flashed back to 1.2 and still could no longer boot into windows so i figured the windows install was somehow corrupt and i'd just reinstall. I gave the bios the benefit of the doubt and decided to give it another try and this time i flashed with awdflash. The flash was nothing like it was with 1.2 something was definately up and all of a sudden the board just rebooted.
It came back on with no video. I turned it off pulled the battery cleared cmos again and it does come on but i get no video. I dont even care to figure out whats wrong with it. The bios definately isnt kosher, i get all the flash at your own risk stuff but i've flashed this board enough times with no issues to know when theres something that just isnt right. Im so done with this board. I dont even want to deal with a rma or even selling it if i can get it back on. I just want to take it in my backyard and set it on fire.
I would recommend to anyone thats flashing to do so from within windows. The bios just didnt funtion the same with awdflash and i believe the other person that had issues was using awdflash also.
this is my 03 score with single x1950xtxQuote:
Originally Posted by coop
http://service.futuremark.com/compare?2k3=4994390
and 05 single aswell
http://service.futuremark.com/compare?3dm05=2253734
but i have'nt run aqua for real i think i have 147k+ with single to..
ALFA bios file with the 1333 strap, didn't help me as for the fsb part.....PLUS it "seems" to work the rams at 5:4 (lower), actually they work 1:1 only.....So no changes and no more fsb.....
A tip for some of you who encode problems after flashing bios or crashes: Put the jumper in crear c-mos pins and REMOVE the battery for about a minute or so.......ALL will come up nice after that....;)
A MUST for more fsb is the Wire Northbridge mod AND the VFSB mod..... ;)
Here is where I'm at. Ran the new prime last night for 7 hours and got no errors. Im at 380 FSB, no matter what I do I can not prime at 400 FSB for more then 10 minutes. Anyone have any ideas?
My max fsb appears to be unchanged with the new bios. Going to try my hand at soldering the vmch and vfsb mods this weekend if I have time. Can anyone confirm if Kentsfield is running all 4 cores higher than 10x with this bios?
Couldn't get ORTHOS blend or the new Prime95 blend to run over 10mins at 401FSB / 1.8vMCH. Looks like the B13 final has nothing special with it.
@pcmoddingmy, have you tried mch at like 1.86 to 1.92v for a prime attempt?
@chefnr1, that is fine 05 score. Do or did you run them (03) with the same settings and speed? Why are our scores just the opposite?
My max FSB is unchanged for Conroe and Kentsfield :(. The good thing about this BIOS is, it didn't break my board like what others are experiencing. In any case, BIOS upgrade always carry a risk so, it's all good.
I have an additional bit of info to add to that. I mentioned that a board I had initially would not start up. It had nothing to do with a bios flash. Just would not boot or even start. Well my present board did the same thing after using it for awhile. I removed/replaced the battery and it was fine. Since it was the same thing that happened to my first board, I removed/replaced the battery and (bios chip) in it and it booted up anew. So I am going to remod my first board. So then I will have two boards and cpu's. One I will get rid of. So if your board gets to a point where it won't start when the on switch is pushed try removing the battery an then replacing it.Quote:
Originally Posted by hipro5
Actually, abit assured me that the board does 400FSB at just 1.75vMCH. I was runnig dual Windows Memtest at 1.85vMCH just now and the NB heatsink got too hot to touch after over 300% coverage in Memtest.Quote:
Originally Posted by coop
Now pumped it up to 1.95vMCH and running Prime95v251 at 401FSB.
From the way I look at it, the heatpipe cooling on the AW9D-MAX is not good, especially the NB heatsink/heapipe. The SB heatsink/heapipe is better than the NB heatsink/heapipe because the SB uses a copper core. Another thing I didn't like about the NB heatsink is that the heatpipe that connects to the rear PWM heatsink is not soldered to it! :nono:
So, heat from the NB heatpipe is not getting efficiently transferred to the PWM heatsink. I bet if the heatpipes were properly soldered to the PWM heatsink, the PWM heatsink should get as hot as the NB heatsink, or at least a little warmer. Now the PWM heatsink is just cool...
Hi all, finally read thru the thread, but I'd appreciate a little bit of help.
I'm stuck at the 9*375 wall. I'm looking for a stable 24/7 use overclock, not just a super pi bench, but I'm not afraid of a little voltage, I generally upgrade every 6 months. I feel like I ought to be able to achieve more from this system as I'm good with stock voltages all the way to 9*370!
Here's the hardware:
e6600 week 29
AW9D Max bios 13 beta 3
Corsair HX 620W
Gecube x1900xtx
WD 150gb Raptor
Cooled by:
cpu-lapped TTBT with a Delta 120x76mm fan (typically keeps it under 45c load)
northbridge-socket 370 aluminum heatsink--roughly 1.5x size the original cooler + 50mm fan
AS5 replaced under PWM block and southbridge
Ram: Crucial Anniversary 2x1gb, Crucial Ballistix PC8000 2x1gb
Here's my approach to 9*375
Crucial Anniversary sticks:
9*371 memory at 1:1, 3-3-3-12-35
CPU voltage (bios settings) 1.2375V, MCH 1.52, RAM 2.2V
This is utterly stable, and my usual setting. Folding 24.7 for several weeks now.
9*375 memory at 1:1, 3-3-3-12-35
CPU voltage 1.2625, MCH 1.64V, RAM 2.25V
Orthos blend stable 1 hour.
9*380 orthos unstable.
I've tried 3 notches on the cpu core voltage, MCH all the way to 1.85V, Ram to 5-5-5-15-42, and Ram voltage to 2.35V. I can't get it more than 10s orthos blend stable.
I've tried dropping the multiplier to 6x, same wall.
In 4:5, I can achieve:
9*316 memory at 4:5 or 395mhz, 3-3-3-12-35
CPU voltage 1.2625, MCH 1.52V, Ram 2.3V
or
9*340 memory at 4:5 or 427mhz 5-5-5-15-35
CPU voltage 1.2625, MCH 1.55V, RAM 2.35V
These were my max's 1 hour orthos stable.
Thinking, maybe its my RAM holding me back, I sprung for the Crucial Ballistix PC2-8000 for some RAM that is qualified for these speeds. And guess what, they are WORSE than my Anniversary sticks. This was my maximum overclock:
Crucial Ballistix DDR2-8000
9*340 memory at 1:1 5-5-5-15-35
CPU 1.2375, MCH 1.55V, Ram 2.35V
Oddly, enough, the Ballistix stix performed BETTER at 4:5
9*350 memory at 4:5 5-5-5-15-35
CPU 1.2375, MCH 1.6V, RAM 2.3V
Again 1 hour orthos blend stable
So to summarize:
Ideally I'm looking to get past the 9*375 1:1 wall for a stable use overclock. I've got 2 sets of RAM that ought to be able to do it easily, by rating or by stable operation at faster speeds in 4:5. I've got a cpu that's still hovering at stock voltages. I'm in the range requiring increasing MCH voltage, but 1.85V didn't help at all.
I understand orthos blend is coming under question, but at 9*380, it is unstable. Orthos only seconds. Will not pass 3dmark05.
And WTF is up with my Ballstix RAM? Can't do 340mhz in 1:1?
Anyone have any thoughts? I greatly appreciate it!
Future plans: replace the nb with the ol' copper dynatron XP, as well as the southbridge with something else.
VFSB mod? Could someone help me post the link to the Vfsb mod?
Thanks,
navig
@Navig, read back a few post for a vfsb mod that can be done without soldering. Another thread has soldering pics in the modding section. http://www.xtremesystems.org/forums/...d.php?t=118248You have come across something I mentioned also in an earlier post about running the mems at 1:1. I found too that if I run 4:5 I can run my rams pretty high but 1:1 requires alot more voltage. Considerably more than man. specs. for the same speed. For example, at 4:5 only 2.4 volts required for 490. But for 470+ it takes 2.65v or more using 1:1 ratio. I don't know if anyone else has this issue or just me.
This board needs an aftermarket chipset cooling before it even powers up... :p:Quote:
Originally Posted by pcmoddingmy
yea, i use 1.75v for 400FSB 24/7 and its very stable...Quote:
Originally Posted by pcmoddingmy
I got my system up and runnig. So far so good. It puts my 3gig P4 to shame. I still need to tweak it. Will see what I can get it to do.
Navig, have you tried some serious vcore? 1.2375v, and even 3 notches up still doesn't get the cpu the juice it needs. up to certain point 6600s don't need much but once you've reached it, it takes more than "3 notches" to climb up. you might want to consider setting it somewhere between 1.4 and 1.5v actual.
yes they are runned at about the same speed only a little change on the gfx core a few mhz on 05, and what do you mean by "Why are our scores just the opposite"Quote:
Originally Posted by coop
Yup... I had a Thermalright HR-05 installed on the NB before. Even 2.0vMCH was not ORTHOS/Prime95v251 Blend stable.Quote:
Originally Posted by JoeBar
1.75V at 400FSB is good for 24/7, runs everything you throw at it, 3DMark, games, dual Spi 32M and all... but when it comes to ORTHOS or Prime95 blend test, it ERRORs out! :stick:Quote:
Originally Posted by Budwise
Quote:
Originally Posted by pcmoddingmy
i have same problem with my P5W64 WS asus:slapass:
my ram will run 450mhz (ddr900) @ 4.4.4.12
2.1 volts @ stock CPU 266FSB.
if i raise the FSB to 401, ram @ ddr800 it will fail orthos blend
within seconds i havent tried prime yet
could this be a bad batch of intel chipsets?
Do not run blend then, run small FFT.Quote:
Originally Posted by pcmoddingmy
@chefnr1, my 03 was higher but your 05 was considerably higher than mine. That is why I asked. I ran mine at the same speeds/ settings for both.
I got my other board finished last night.
definitely a bios issue with blend and memtest.
i tried testing my supertalents at 4-4-4-12 @ 420 mhz and i couldn't do blend or memtest (dos version). i tried every combination of voltages known to man.
then i tested the mem on my p5b... at 2.25 volts i can do 4-4-4-10 @ 475 with no errors for 24/7 use.
it has to be the bios... i think above 400 mhz the NB timings might be super tight... hopefully a bios update will fix it...
For me, I had good results with P5W64 WS PRO. It needed only 1.80vMCH to do 429FSB, and it did over 14 hours ORTHOS Blend.Quote:
Originally Posted by nanoprobe32
From what I'm seeing, its not the chipset being bad batch. Almost everyone here is having Blend issues with AW9D-MAX. The board runs fine on everything else, except Blend. If the board runs blend test for days on 266FSB, it should do the same thing at 400FSB. If it fails Blend, its not totally stable in my book.
Btw, running Small FFTs is doesn't prove anything... DUAL Windows Memtest or Blend is the best way to determine system stability. :fact: