Which DFI NF4 bios for Twinmos AA4T UTT?

[Monge]

Hey folks.

Which bios for DFI NF4 Ultra-D will give the highest memory overclock with 2*512MB Twinmos Speed Premium PC3200 on Venice 3000+ 0517DPMW?
I'm on N4D510-2 bios now and 260MHz@3.3v 2-2-2-5 Superpi 32M stable.

Monge

[Major]

the 510-2 is going to be hard to beat with UTT, if I were going to try another it would be 704-2 bigtoe mod but I don't think your going to see any change in ram OC.

[Monge]

the 510-2 is going to be hard to beat with UTT, if I were going to try another it would be 704-2 bigtoe mod but I don't think your going to see any change in ram OC.

Major: Thanks
I tried 704-2BT but I gained no more memory MHz. Did you try 510-2 fix or Merlin and Helfire modded bios?

Monge

[JfRsQ]

Depending of which ram divider you will use, 9/10 seems to be unworkable passed bios 510-2.
Each time I'm trying to use this divider with later bios than 510, the screen is freezing at the Backup CMOS part after the DMI information and can't boot into windows @2750Mhz (275x10) DDR500 ram with the 9/10 multi but I can with @3000Mhz (300x10) DDR500 with the 5/6 multi.

Also, I moved to 704-2 Bigtoe bios 2 days ago and I'm getting a new BSOD callled Trap_Cause while running @3Ghz. I googled this error and found only 3 webpages about it.... all in german

I will give a try to Xcel and merlin bios. If they are not better, I'm going back on the original 510-2 or fixed bios and will camp it until I'm done with A64 and DFI

Update:

Trap_cause seems to be gone after switching both ram to the YELLOW slots and 3 bios (bigtoe, Merlin, Xcel2updated) where I was getting this error in Everest Writing Test.

[Monge]

Little update: I'm trying the 510-2FIX now and I saw no increase in memory speed. It seem to be the same.
Also I did try the 510-2 Merlin one and saw no increase in memory speed with my Twinmos SP3200 UTT/CH-5. Maybe I'm hitting a wall with my memory or it's my Venice memory controller, that's freaking out.
btw I'm using the orange memory slots - better memoryoverclock than yellow on my board.

[higgins]

using 623-2, tried 510-2 merlin... dont see any real diff between em. but on the other hand im just doin 2228 230 3.2V with my sp3500.. wont touch the 5V jumper... no cash to replace my mem if it fries

[Aleman]

using 623-2, tried 510-2 merlin... dont see any real diff between em. but on the other hand im just doin 2228 230 3.2V with my sp3500.. wont touch the 5V jumper... no cash to replace my mem if it fries

Using >3.1V on the 3.3V aint healthy anyway!

[higgins]

Using >3.1V on the 3.3V aint healthy anyway!

really? why is that?

[Monge]

really? why is that?

Cause your 3.3v rail has to be +200mv or +0.2v over actual vdimm value.
To get vdimm = 3.2v you need 3.4v on the 3.3v rail.
0.2 - 025v overhead on the rail feeding the memory is preferable.

Look here:
http://www.xtremesystems.org/forums/...692#post851692

[Monge]

I'm using 510-2 by Hellfire now and I can say I gained a little more overclock with my Venice 3000+ DPMW than with bios 510-2 and 510-2Fix. Memory's about the same, but I have to test the Hellfire bios more in the next days. But my first impression is very possitive.