DFI SLI-D & Opteron 146 RAM Issues

Jimbo Mahoney · 21 Jan 2006 at 11:26

TBirdUK said:
Been to the DFI-street website and cant find the settings for the Geil Value dual channel 2 x 1 GB. Machine shut itself down again last night while blend test!

Have run memtest for 7 hours and no errors. Think it could be a driver issue? or any ideas?

Did you try my settings?

I found memtest to be a VERY poor indicator of system or memory stability. It just gives you an idea.

Using Small FFT or Large FFT (NOT blend) in Prime should help you narrow the problem down a bit. I always found Large FFT harder to pass because (I suspect) the memory controller is the weakest part of the chip.

TBirdUK · 22 Jan 2006 at 14:29

Jimbo, yep tried them but as the problem is intermittant the testing can take anything upto 24 hours. Seems to work with small and large but blend is still a shutdown @ around 81/2 hours.

TBirdUK · 2 Feb 2006 at 22:24

Darn it, changed the PSU and still when the machine is under load it just shuts off!!!!!!

Anyone else had this problem??
Temps are good, memtest works, do you think it could be a graphics card issue?? (6800)

Help please I've run out of ideas!

ajgoodfellow · 3 Feb 2006 at 02:14

Have you tried giving your CPU more volts?

The memory controller is on the CPU so with your heavily overclocked CPU it might need more volts for the memory controller to remain stable when accessing a large amount of memory

Big.Wayne · 3 Feb 2006 at 02:45

TBirdUK said:
Got the 146 running at 3Ghz but the RAM will only boot with a dividerof 100(1/02) with very slack timings and 2T

Hi,

I have the exact same issue, Opteron 146 with ASUS A8N-SLI Premium amd 1GB(2x512) of Ballistix PC3200.

I suspect its the memory not liking dividers, happens sometimes. After a while I decided to change the memory back to 1:1 sync and have managed to push it up to 275HTT, it would even do 250HTT when I was using a 5:4 memory divider :confused:

As soon as I finish Priming my memory at 275MHz 3-3-3-8 (1T) I will put in some other sticks and see if this divider issue vanishes! (hope so, otherwise it may be something to do with the on-die memory controller!! :eek:

).

TBirdUK · 3 Feb 2006 at 09:10

Thanks guys, whats the max volts for a 146, I'm already up to 1.66 (1.55*113%) and am slightly worried about pushing it. Rig is on water and temps peak at 49 under load. Would the memory controller cause a shutoff?

Big.Wayne - I've got basic memory and it wont go above 225, doesnt even like that very much (but I have memtested and it passes) so a 1-1 issue is not good, however I dropped the core to 2500 last night and it worked for ages. Trouble is the shutting down is so random - it has primed/BF2 for anything from 10 mins to 8 hours, so testing the problem can take a lot of time.

Big.Wayne · 3 Feb 2006 at 13:48

TBirdUK said:
Thanks guys, whats the max volts for a 146, I'm already up to 1.66 (1.55*113%) and am slightly worried about pushing it

Hi,

I don't think vCore is the answer here, I followed a similar path when I first attempted the Opteron overclocks, but after getting in a muddle with this non-working memory divider thingy, i scrapped everything and started again.

Whereas before I was needing 1.55vCore to try and get things stable while overclocking the Opteron on a divider, now I am runing at stock volts (1.4v) and the CPU is up to 2.75GHz (275HTT X10, Memory 1:1 sync) and rock solid. Either I got a good chip or most Opterons would seem to be *very* good silicon.

Patience is the key here I think, it's my first A64/overclock so Im just taking it slowly and trying to make some sense of it!.

Have you got access to any other Memory sticks that deffo work on a divider? That would be worth a shot! . . . .

TBirdUK · 14 Feb 2006 at 07:40

Changed the PSU/mobo, got it working on 2700 on air, good temps. Think there was an issue with the mobo/psu. Have a 165 in that mobo with a new psu and no probles. Thanks for all the advice - overall learnt a lot - think it was a DFI compatability issue overall.