Help with getting the best from my system (updated and solved?)

I meant Prime95 yeah.
Bit suprised with it being stock clock.

 

Still unstable with all voltages set to stock - but actually everything but the vCore already was. The NB voltage was set to 1.5V (the lowest it allows), but is being recorded as 1.7something. The vDimm always seems to be about .2v above the setting too...

 

Ok cool, probs just read errors then perhaps...

Maybe try setting bypass max to 7 or 8, see if that helps. Also, maybe set the memory on a 166/180 divider @ 250 HTT w/9* cpu multi and see if its stable there...
Could be htt speed perhaps, rather than the ram

 

I've just tried setting MEMCLK to 166 in A64 Tweaker, and everything is stable at that speed (204MHz, apparently). At any higher memclk settings, Prime95 fails within 2 minutes.

 

Ok, so its sounding like the ram *might* be the issue here.
If you boot at 250 1:1 (mem in sync with HTT) and run memtest86+ how do things look?
You just need test 5, give it 15-20 loops.

Can also run "memtest for windows" to check things out in windows, its probably worth giving this a go too tbh...

 

Memtest86 ran 25 times with no errors, but memtest for windows showed 18 errors in the first minute!

At stock speeds, it doesn't find any errors.

 

Ok, thats interesting.
Memtest for windows stresses the system as a whole a little more than memtest86 does.

Found alpha timings helped quite a lot getting windows stable, its kinda a suck it and see affair with those tbh, don't know how fully featured your bios is either, what sort of contol do you have timing wise?

Would try setting tras to 10 or 11 and see what happens. Do you have options like tref (refresh) drive strength/bypass/idle timer in bios?

 

I tried tras 11, but it made no difference. I dom't think those options are in the bios, but i'll check to make sure.

__

Nothing there that i can see!

 

Gah! Not what we needed to hear!
erm, ok... *thinks* the memory *should* be stable at 250 with the timings given. We know the board and CPU are happy with the HTT, running the 166 div showed us that.

We know the rams ok at 200 (under stock but still) Have you tried clearing cmos n the usual stuff, force all defaults n see what happens then?

 

Not sure if this has been suggested or what the problem really is (as I only scanned the thread) but my redlines need a lot of voltage to get to 250MHz with timings of 1.5,2,2,7 1T... My DFI SLI-D is maxed out (without changing the jumper setting as it is know to kill ram) which if I remember right is 3.2v.

 

His redline doesnt use winbond UTT like yours does.

Just had a quick thought, whip out the X-Fi and see if that helps

 

Cleared CMOS, and no change (once i'd set the HTT back to 250 etc)

Took the X-Fi out too, and it made no difference. I was kinda glad, as if it had worked it would have been embarassing that I hadn't thought of it before

Is it about time I started on the long harrowing journey for an RMA, do you think? I don't have access to any other systems to try it in unfortunately.

 

Not sure tbh, do you deff have no way of testing the ram elsewhere?

 

No easy way that i can think of :/

I just finished testing each stick individually in this system, and neither give any errors when on their own.

 

Even at 250?

Ok, try swapping the order of the sticks over and see if that helps.
Also try switching to the other memory slots and see if they give you any better reults
(ie if you were using 1&2 before, use 3&4 now)

 

Yeah, both were fine alone at 250.

Swapping them over didn't seem to make any difference... I'll try again incase i swapped them over twice in one go

When i put them in 3&4, I got a BSOD as soon as I reached Windows.

I'll try both again in the morning, when I'm a bit more awake!

 

UPDATE :

My new DFI lanparty UT SLI-DR Expert (why do the names have to be so long?!) arrived yesterday, and the RAM is now working fine at its rated speed. Must have just been the crappy fatal1ty board... Damn you Wendell!

Of course now SLI isn't working

Second update :

Got the SLI working... I see where this thing got its name from... that wasn't exactly plug and play!

 

(Apologies if this has already been posted, I quickly scanned the thread)

Doesn't redline need 3.2/3.3V to run at it's rated speeds?

 

IIRC some of the older stuff did, but the 2GB kits use Infineon CE-6s, which run well with 2.7-2.9V.

 

The redline used here isnt the UTT chips, thus it doesnt require high voltages.