• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Memory/ Stability problems

V

verdstein

Junior Member
Last week, I finished building a computer (to be used mostly for scientific applications). When I install 4 DIMMS (2 Gbyte) of memory it is unstable. It freezes or reboots about once per hour during normal operation. I can always make it freeze or reboot by opening ~10, 5Mbyte video files in rapid succession. The system also crashed during a Goldmemory memory test run from a floppy disk at startup.

I tried installing all the DIMMs in different combinations. I found that with 2 sticks in dual channel mode, the system is always stable, no matter which sticks i use, and whether I install them in Channel A or Channel B.

The memory voltage is currently set at 2.6V.

The system runs cool. CPU temp ranges from 25-40 C and case temperature is 25-35 C.

Any ideas what might cause a problem like this, or what else I can do to diagnose or fix the problem?

Motherboard: Gigabyte GA-K8N Ultra-9
VGA Brand : Gigabyte GeForce 6600 GV-NX66256DP
CPU Brand : AMD 64 3800+ Rev. E Venice
CPU Heatsink: Thermaltake xp-120, ArcticSiver Ceramique, 12 cm fan
Operating System : Win XP sp2
Memory : 4 DIMMS Crucial CT6464Z40B DDR400 512 Mbyte
Power Supply : Enermax Noisetaker 480 W

 
Simply put, just bad RAM module(s).

Not unusual or rare to get these days.
Determining which module it is might be a problem.
 
What speed is the memory running at when you have 4 sticks in? What core is your A-64? If it's not a venice or san diego, you'll need to have your ram at 333mhz at 2T with 4 sticks.
 
If it is a bad ram module, could I test them one at a time or in pairs to see which one reproduces the instability?

Memory is running at 400Mhz. My cpu is a Athlon 64 3800+ 512K 90nm Rev. E Venice, so that should be OK, right?
 
Try increasing the RAM voltage to 2.7. Did you try a software monitoring utility to check the actual DDR voltage? Also, I'd test one module at a time using the same procedure that crashes the system. You might be able to quickly pinpoint a bad module that way.
 
Thanks BadThad. I increased the voltage to 2.7. Same problem. I checked the voltage in EasyTuner, (which comes with the motherboard). When I also checked the voltage with SmartFan, I saw that it listed the voltage at 2.64 instead of 2.7. I bumped it up again to 2.8 (2.74 as reported by EasyTuner). There was no improvement.

I tested each module separately. None of the individual modules reproduced the problem that I experience when I use all 4 at once. I've also tested all of them in pairs, in channels A and B, and all the two-module combinations were stable.

Gigabyte tech support has no idea what is going on.

Any other ideas?

Maybe I should just live with 2 modules (which are sufficient for 98% of what I do) and not waste any more time on the problem, but somehow that feels like I've been defeated. 😛
 
There still might be timing issues specific to your motherboard. Try running all four DIMMs at 166Mhz and 2T timings. Use memtest86+ (www.memtest.org) to see if any errors occur when you test.
 
Good call. With all four DIMMS at 166Mhz and 2T timings, the system appears to be stable. I haven't had time yet to run memtest86+, but the obvious instabilities that were present before (e.g. freezing every single time I open 10 video files in rapid succession) are gone.

So, what does this mean? Are there settings I can tweak to get this memory to work at 200Mhz? Would a different brand/ model of memory run with 4 DIMMs at 200Mhz?

Is this a compatibility issue between this particular model of memory with this motherboard? . Bios automatically set this memory to 200 Mhz. I can't find any mention in the motherboard manual that the memory clock speed must be set to 166 when using 4 modules. Also, I told Gigabyte tech support that I had 4 DIMMs running at 200 Mhz, and they did not tell me to change the memory clock speed. It seems to me that this motherboard is supposed to be able to run 4 DIMMs at 200Mhz.
 
the problem is with the memory controller on the chip, cant believe gigabyte didnt know that. only the a64s with the venice core could support 4 dimms at 200mhz.
 
My cpu does have a Venice core. A lot of good it's doing me. :disgust:

So, my question is still whether I can hope to achieve a stable system with 4 modules running at 200Mhz, or whether the timing issues specific to this motherboard will prevent it no matter how I tweak it and what brand/ model of memory I use. I've been scrounging for this sort of info for the GA K8N-Ultra-9 and I haven't found anything.
 
Originally posted by: verdstein
Good call. With all four DIMMS at 166Mhz and 2T timings, the system appears to be stable. I haven't had time yet to run memtest86+, but the obvious instabilities that were present before (e.g. freezing every single time I open 10 video files in rapid succession) are gone.

So, what does this mean? Are there settings I can tweak to get this memory to work at 200Mhz? Would a different brand/ model of memory run with 4 DIMMs at 200Mhz?

Is this a compatibility issue between this particular model of memory with this motherboard? . Bios automatically set this memory to 200 Mhz. I can't find any mention in the motherboard manual that the memory clock speed must be set to 166 when using 4 modules. Also, I told Gigabyte tech support that I had 4 DIMMs running at 200 Mhz, and they did not tell me to change the memory clock speed. It seems to me that this motherboard is supposed to be able to run 4 DIMMs at 200Mhz.
Click to expand...

Beyond pumping a lot of voltage across the bus, I don't think there's any settings you could tweak that would make a difference, although I would appreciate if somebody would prove me wrong, because I have a similiar problem. I'm no EE, but I would imagine that since the AMD BIOS and Kernel Developer's Guide recommends that a Socket 939 motherboard design only be validated with 4 double rank DIMMs at 166Mhz with a 2 clock bus turnaround, there's some electrical issue inherit to it's design that prevents it from running at 200Mhz 1 clock bus turnaround, making future CPU support for it a moot point.

http://www.amd.com/us-en/assets/content...e/white_papers_and_tech_docs/26094.PDF
See section 4 DRAM Configuration.

It could also be an issue with the particular DIMMs you're using, for example they could have a poor PCB or chip design that puts alot of noise onto the data bus, or they have an out of spec impedience, or a thousand things, enough to cause errors when using 4 DIMMs at 200Mhz. But the only way to tell for sure is to try some different ram or a different board, of course.
 
Joepublic, thanks for your thorough help! I guess I may have to subdue my compulsion to fill all slots on the motherboard. 😛
BTW, I did run memtest86 eventually with all 4 DIMMs at 166 Mhz and they passed.
I'll give an update if I make any discoveries or get any more info from Gigabyte.
 
MEMTEST.

Yeah, thats the ticket....

While memtest is better than say , nothing, there is no better test then a
hardware ram tester.

Just because it passed memtest once does not mean it will the next time.
Run it for yourself, and few days apart. Guaranteed different results.
 
Good point. What to do about it, I'm not sure. You think it's worth the time/ $ to track down a hardware ram tester, or was it more a rhetorical comment?

edit: I called a couple of professional repair techs. They didn't have one and didn't know who would. I'm guessing this isn't going to be worth the time.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top