Is this maintainance menu part of the new bios or is this something that you only see when the bios reset jumper is on certain pins? When flashing from 2333 with an o/c to 3.2ghz on a 6600 will i need to reset all settings to stock before the flash to 2507 and then disable this watchdog thing? If i don't go to stock will watchdog keep me from posting after the flash? What method did you guys all use to flash to 2507 and return to your original o/c?Quote:
Originally Posted by eagle101
Well I don't know is there any bios problems with this mobo and quad CPU, I haven't found any information besides the "reviews" about OC quads with this mobo, but my little experience may be help others.
It took me a while to stabilize my rig, I've tried many settings and really don't know if the quads stresses big time the chipset because with 266 strap The highest I could go stable was 380*8=3040 mhz with 1.7 v MCH and 1.5 v FSB :confused: and with strap 333 395*8=3160 even with memory at 400 I couldn't raise the bus speed any further :mad: .
I've disabled EIST, C1 and watchdog, Disabling VT and XD didn't improve the OC.
I've tried the 2333 bios but things got worse so I reflashed to 2507. :(
With respect to the temperatures I've a 170 mm fan on the side cover of the case that pushes a lot of air and keeps cool my R1900XTX, NB, SB and ram. :cool:
Any way I'll keep trying to Oc higher.
Cheers to every one
You can only see the Manintenance Menu when the BIOS jumper is in the Configure position. Try this:Quote:
Originally Posted by astrodave
1) Remove power to power supply by pulling the cord or switching it off.
2) Wait ~5 sec or until the light goes out on the M/B and move BIOS jumper from the Normal pos (pins 1-2) to Configure pos (pins 2-3).
3) Restart PC and the Maintenance Menu will be displayed after POST.
I think it's always a good idea to reset to stock clocks, not necessarily stock voltages, before flashing to ensure 1000% stability ((not that I always practice what I preach ;) ) . You're right that Watchdog may prevent you from going into the BIOS if you flash from a previously OC'ed state - but you can pretty much almost always get into the Maintenance Menu by switching the jumper to correct any problems.
Maybe try a little higher on the FSB voltage? I ran my CPU at 1.4 and the FSB on the next lower voltage (at work, dont have the stock V#) and that was stable for an hour. Lowering my Vcore to where it's at now with the FSB still stock crashed Orthos. Bumping the FSB up made it stable at the lower Vcore (6 hours overnight). What I'm trying to say maybe higher FSB will let you run lower Vcore.Quote:
Originally Posted by Eluzion
I see you have a Tuniq in your sig. What temp are you seeing in TAT & what RPM are you running the fan at?
I was looking into this, I think I'll stay stock with AS5 treatment and slap a 40mm fan on that (the one Swiftec uses on their chipset cooler looks interesting). The stock heatsink is a big hunk of metal, so maybe it is just fine and you could add a fan if you are running a tower cooler that doesnt create fan wash off a board mounted CPU cooling fan, for piece of mind.Quote:
Originally Posted by casse
The IDCC windows app locked up when trying to adjust board operational things in there for me, so I just stick with bios. Seems that thing is only good for adjusting the fan header settings.
I got a question for those of you that have played with using decent memory on these boards. I'm using the 2333 bios, and 3.6GHz/400FSBx9, 1:1 266/533 memory to make 800ddr2. Right now I have some crappy Promos Corsair 6400C4 4-4-4-12 running. It wont go much higher than where it is at now.
I am thinking of getting some Ballistix D9 based memory and running it either at 1:1 266/533 with 3-3-3-9 or so timings, or trying to go to the 200 strap and see what I can do there (the 'unseen' timings are tighter down there maybe?). Any thoughts if this would be worthwhile or successful? I heard that going beyond 1:1 on a 975x chipset didnt have the same gains as like on a 965 doing 4:5, so I want to explore tighter timings. What have you guys found?
With IDCC I have found that trying to dynamically change settings in windows will never work, but if you have it save and reboot it will apply the same settings correctly. Pretty much just a prettier interface to bios overclocking.
--------Watchdog Timer---------
- Power down the system and pull the power cord
- Hold down the power button on your case to discharge the PSU for 5 sec or so
- Remove the BIOS jumper
- Plug power back in and reboot into BIOS
- Watchdog timer is under the very first tab that appears, shut it off, along with all the other unneccessary stuff
- You prompted to power down and replace the jumper, do so now
- Now overclock to your heart's content ;)
Yeah, I've seen this. There are definitely ref/mem frequency combinations that just don't work and I found several cases where changing the combination worked even if it resulted in a higher actual memory frequency.Quote:
Originally Posted by Omastar
The only suggestion I can make is to keep trying various combinations.
IDCC works well for me when adjusting FSB. it applies the requested FSB and CPUz and so on verify it.
however I randomly cant get IDCC to initialize and usually after waiting 10minutes it starts noproblem.
I have experienced the same thing as far as memory
If I push the 1066 Strap too much (past 398 or so) and have the 667 memory divider, my memory gets errors immediately using the windows version of memtest. If I use 400FSB. I will get errors instantly. I have even tried setting my mem volts to 2.45v with no luck. 1000MHz is impossible for me with 1066 strap, yet 970MHz runs completely fine with 2.16v.
I know the Northbridge / memory controller is at fault here, because if i increase the strap to 1333, I can memtest fine at 1000MHz memory. Increase vMCH and vFSB did not seem to help memory stability at all on the 1066Strap.
Small question:
I use the intel badaxe 2 board, wich has 8 sata port. 1 on marvel connector (blue sata) 1 on intel matrix (black sata).
I have one computer setup with 4 seagate 320gb in marvel sata and 1 raptor in matrix.
the other computer i want to try something to have more performance. 4x320gb on each controller, and a system disk on IDE.
Do you think 2 raid 0 (4 disk each) would give better performance?
The system will be used for heavy HD Compositing.
Thank you
2 raid0's with 4 drives each is going to fly. Make sure you select the 128kb stripe size and force 64k allocation units when you format the drive. The Intel will edge out the Marvell on burst speed, but from a thoughput point of view they'll be pretty close to each other.Quote:
Originally Posted by shogo_ca
You've got me thinking now... :)
Quote:
Originally Posted by GPSeek
It's efficient enough and really quiet and has enough power for most rigs. But they have a high fail rate. Antec don't really have a QC. But if it doesn't break. There is no reason to switch. "if it aint broken don't fix it".
You may be right with the efficiency. However, if this Antec line of products has a high failure rate, you're risking your new system. If only the PSU dies, that's fine. But it may burn your other components. :slapass:Quote:
Originally Posted by Poodle
Cool Cool im gonna try that tonight. For the other array that were created do you think its worth deleting them and create with the same settings (64k allocation, 128k stripe) because im not sure i used those settings, i just kept default....Quote:
Originally Posted by gtj
I have the same thing happening but...using the 10x multi all is good at 400fsb 1066strap 667 divider. Using 9x and 8x i get mem errors. Very wierd"?Quote:
Originally Posted by Miwo
I can't get my e6600 stable at 3.4ghz. It fails orthos with the expected rounding error. My chip is an L629B367. It passes windows memtest.
Settings:
vcore 1.45
vmch 1.6
vfsb 1.325
fsb 380
ram at 533 4-4-4-12 2.16v
c1e, watchdog, eist disabled
enhanced power slope enabled
Temps are about 45C load.
Also cpu-z, IDU, etc all report my vcore as about 1.325 no matter what I set it to in the bios. The hardware monitoring page in the bios gives me 00000 for all the values as well.
Any suggestions, or is my chip just maxed out?
I think you'll be ok with the defaults. I think the stripe size defaults to 128 anyway.Quote:
Originally Posted by shogo_ca
http://img.photobucket.com/albums/v4...dplant72/3.png
My best so far, although I'm not done finding my max overclock yet. RAM doesn't wanna do over DDR2-1000 on the 1066 strap, so I'm at DDR2-800, 4-4-4-8 for now. 1.45vcore, could likely do 1.4v.
My thoughts on the XBX2...
Intel needs to get their manufacturing process better on these boards. I'm guessing it is the chipset, but it could be power delivery or decoupling cap issues which are leading to such varied results. When I first got the board worked great fine and couldnt be happier, but performance gradually got flaky. Now that I've read about 20 or more others with the EXACT same wierd issues, I'm more than convinced these XBX2's have some serious manufacturing flaws and some minor annoying design flaws.
For me, whenever I do a BIOS update regardless if it is from floppy, cd or windows, I get three beep error even though I have reset everything back to default before the flash. I have to remove the CMOS battery AND boot back up with only one memory module.
After that, in order to overclock at all I have to disable the stupid watchdog and all of the other worthless debugging settings plus c1e and eist of course. THEN I must change memory voltage and reboot about 4 to 6 times before it actually sticks in the BIOS. After I add the rest of my memory, I usually get a no post and have to go through the process all again. And the recovery BIOS thingy dont work worth a crap. Basically I keep screwing with it till it boots up 3 hours to 1 day later.
I used to run folding and those kinda programs 24/7 and originally OC'd at 3.85Ghz 100% ORTHOS Blend stable for 12 hours with 1.65v MCH and 1.4Vcore so I did stress the crap outta my system. After a few days I started getting lock ups and thought it was my memory. Well its not memory, its the motherboard that is slowly degrading in performance. I'm now down to 3.7Ghz and it is pretty stable now that I did Vmod for MCH and boosted it to 1.7 volts actual measured.
Also my board wont do 1333 strap any longer at all. My system wont post and I have to go through the day long process of getting it to boot up again...
I think it's pretty poor that about half the people with xbx2 can OC like a mofo, but half are lucky to get 400FSB. and a high percentage have wierd no post/no video problems on a bad OC or BIOS flash. This is WAY too high of tolerance for a production product especially for Intel. Problems like these indicate a poor manufacturing process.
I will bet you that if your getting less than 400FSB with an E6600 it's because of a poor MCH. Try cooling it better like I did and I bet you'll get a bit more stability and OC. MCH Vmod made a huge difference in stability.
For the non posting issues, your on your own... Everyone seams to have their own little ritual that eventually gets them a boot.
On the positive side, the XBX2 has a great RAID. I get 260MB/S in HDTach for sequential reads and 3.5GB/s for burst.
I think I'm gonna go for an 965 based Mobo after this XBX2 burns out in a month or so. It's steadily declined in stability since I got it. More volts on MCH helps and better cooling on northbridge helps, but after a few days it starts doing the lockups again untill I lower FSB and boost Vmch.
Hm, I'm running 400x9 @ 1.475 (load w/ vdrop), 1.5 (set in bios). I've ran Orthos four times, once at one hour, two hours, five hours, and two hours.
I really don't want to push the vcore anymore than this (1.5 is enough lol). Any other settings I might try playing with or is vcore pretty much it?
I hate all of you running 400x9 with < 1.40 vcore. ;)
You blamed the memory first, now the board. :) How do you come to this verdict?! Have you RMA'd your memory sticks and tried new ones?Quote:
Originally Posted by SPL15
The BX2 Bios is not very user friendly. However, nobody plays with it all the time. It's just fine for an overclocker.
Quote:
Originally Posted by GPSeek
Yea bought more Ram and same exact issues to the T. After doing the mods and cooling the MCH better I got rid of the errors at 415MHZ fsb.
The thing that tipped me off what the fact that running one module of ram at DDR6xx gave me the same exact errors as 4 modules at DDR8xx regardless of memory volts.
The memory is fine, I'm running it at DDR830 and its stable up to DDR1012 with all four modules in at 4,4,4,10 2.2v.
Setup a delta fan over northdridge in my garage for testing and my problems went away. Brought system back inside and problems came back. Used a can of freeze spray to cool MCH and ORTHOS would run fine till I stopped spraying and would fail in about one minute after.
Boosted MCH volts with Vmod and I'm fairly stable again, but I;m still far below the 3.85Ghz I was doing initially. CPU has never seen higher than 60C or more than 1.45Vcore and small FFT runs all day long at 3.835Ghz at 1.4volts CPU still.
I'm just amazed how varied the results are with the XBX2. In my experiance in electronics manufacturing, this kind of variance is due to poor manufacturing. Nothing manufactured as a whole assembly such as a motherboard or video card will have that great of performance variance unless something aint right.
EDIT: I actually wish there were more options in the BIOS. I personally think it is quite weak compared to newer nvidia motherboard BIOSes. Now THATS an overclockers BIOS!!
SPL, if you really hate this board so much, consider an Abit AB9 Quad GT. I hear it does 515FSB on 1066 strap all day long. I'm not gonna argue that this board is annoying as hell sometimes, but there is no perfect motherboard. Look around at all the people having problems with the oft-lauded P5B-Deluxe, and the Abit AW9D-MAX issues, and the P5W-DH haters, and the anti-680i people who claim they burned out 8 sets of RAM on their 680i board. It goes on forever, and the only solution is to try different combinations of settings until it works for you. This is the inherent downfall of overclocking--there are no guarantees.
And I realize now about the non-POSTing issue/no video after bad BIOS settings, you should be able to set jumper to pins 2-3 (Config) and change the erroneous setting(s), set the jumper to 1-2 and boot. It's always worked for me. Never had to clear CMOS or anything so far. I think, really, the problem with these boards lies not in its supposed sub-par manufacturing process but moreso in its still immature BIOS. For example, why won't it clock my Ballistix past 1000MHz with any timings/volts? Numerous people have had this problem and the only way around it is to try DDR2-1000+ on the 1333 strap, try a different reference memory frequency (667MHz vs. 800MHz) and hope it works, but there's no coverall solution for every instance. I hope Intel will LISTEN to the knowledgeable people who own this board and take current BIOS shortcomings into consideration for their next BIOS.