Official DFI NF4Ultra-D/SLi owners thread.

mattbrown91 said:
My BIOS is 3/10/05 which is 10th March, i am **** scared from the prospect of flashing the BIOS.

How often do these things go wrong?
Why do they go Wrong and break the board?
They go wrong if you have bad RAM, an unstable system or suffer a power cut during the flash. i.e. rarely but these are the risks. Sometimes just an unlucky bad write will cause it. When they go wrong, they don't break the board, only the CMOS chip. This means that the board cannot boot because it tells the board how to boot and which settings to implement. You can just replace the CMOS chip and there are places online that sell the chip in case this happens (and preload the BIOS you ask them to onto it). They go wrong more often than not in Windows flashing (Winflash) because sometimes a windows thread will interrupt the process causing the problem. DOS flashing is much safer. Also, floppies cause them to go wrong, as floppies are notoriously unreliable. I flash direct from a small FAT32 partition I keep on my hard disk which can be accessed from DOS directly.
 
smids said:
It does but isn't that the one with the Vcore bug, or was that 704? I would just use the 704-2BTA (OCZ Tony's) BIOS. It's a proven BIOS.

Hey Smids,

Hows your ZX now? Mine arrived the other day and I have to say that im well impressed with it. Getting better performance now with tighter timings running at stock speed. :eek: Everything seems to just work better now. Glad I ditched by Crucial sticks. Just goes to show that running ram on a high fsb with slack timings is pointless on the A64 platform. ;)
 
you recon? I've found when I run my ram 3-4-4-8 1:1 with my opty at 2.75gig it beats my super-pi score and encodes videos quicker than when I had my system running at 3ghz and ram ddr400 1t 2-2-2-5 on the 133 divider.
if I could only get my g.skill hz's to a stable 300fsb 1:1 my system would be so damn quick, but thats not at all likely though with these uccc chips, not like tccd.
also, I've noticed if I run oldish games that are slightly cpu limited hi fsb helps quite a bit.
 
Scoobie Dave said:
Hey Smids,

Hows your ZX now? Mine arrived the other day and I have to say that im well impressed with it. Getting better performance now with tighter timings running at stock speed. :eek: Everything seems to just work better now. Glad I ditched by Crucial sticks. Just goes to show that running ram on a high fsb with slack timings is pointless on the A64 platform. ;)
Glad you're happy. Mine's still in RMA at the moment. If it comes back clean (in their testing), then I fear my memory controller can't handle the 2GB of RAM at stock (lockups occured at stock), in which case I'll have to RMA my CPU because putting the 1GB (2x512) sticks in didn't cause any crashing with no settings changed.
 
I noticed cpu's with a weak/picky memory controller or slight incompatibility issues tended to pass memtest but would sporadically lock-up in windows, or programs would hang and in some cases it would bsod when pushed. did you try upping the vcore or knocking the multi down but sustaining the same ram speed?
my cpu no longer likes high dividers with this 2gb g.skill ram now, but will be happy running the ram at much higher 1:1 frequencies.
 
No, I'm not upping the Vcore to run at stock - might as well get a later stepping which probably has a stronger controller. If they contest me, bah, I'll quote as many statutes necessary to get a replacement (that is if it is proven the RMA'ed RAM was not faulty).
 
No m8 I didn't mean it like that, I meant so you can rule out the offending component, without the uncertainty.
No your quite right, you shouldn't have to up stock voltages to run stock speeds.
 
Hesky82 said:
No m8 I didn't mean it like that, I meant so you can rule out the offending component, without the uncertainty.
No your quite right, you shouldn't have to up stock voltages to run stock speeds.
Sorry, I realise I accidentally wrote too strongly and didn't include smileys ;). It's true, I could have done that, but it was obvious from the fact that the moment I put my Geil One back in (2x512MB) then it didn't lockup at all. That instantly made me suspect the memory controller. I've contacted the seller anyway, just in case the RAM wasn't faulty (and charge them for any testing fees I incur - though that may require a bit of reading up on my part).
 
smids said:
Good CD - and Tmod added so many useful things on that CD as well as implementing memtest86+ v 1.65 into each BIOS.
I have spent the last hour looking and threads on DFI street to find out if the 623-3 BIOS or 704-2BTA BIOS is best with Ballistix memory but did not find any info on that.
I think I will try the 704-2BTA version and see if that gives me better stability when using a memory divider.
 
Back
Top Bottom