Is it my board of my RAM?

my specs are in my sig. I used to run my E6600 at 3.6 or 3.2 ghz. Note I have 4x1gb of ram. I eventually reduced clocks to stock because I started getting some BSODs. This made it better but I still have periodic issues. I am certain that either my NB is shot, one DIMM is shot, or 1 or more ram modules is shot.

Over the summer, I ran Memtest with all four dimms filled and got errors. I then ran one module and it passed. I subsequently ran the cleared module through each DIMM and they all cleared. My guess is the NB cant handle managing four dimms at once, even at increase voltage.

Do you guys know of another way to definitively show this? As well, can you guys recommend a good overclocking replacement board that is fairly cheap?

Thanks,
Phil

This happens once in a while.

1) Is the RAM on the motherboard support list?
2) How's the power supply? I've seen bad PSUs cause errors on loaded-up motherboards during RAM-intensive tests. Swap/test if you can.
3) Try two modules at a time.
4) Try four different modules if you can get hold of some. Of course, you may not be able to - I'm lucky in that I have access to all sorts of this stuff at work. Most don't.

I've run into the full-RAM slot problem lately. I have an old AOpen AK77-600n motherboard for the Athlon XP. It was specced with 3 RAM slots, with a theoretical max of 3GB of RAM. So, I tried to load it with 3 Mushkin 1GB modules, and got a whole slough of errors. Tested each one individually - fine. Tested 2 at a time in every combination of slots possible - fine. Tested with 3 modules again - fail. Test with 3 512MB modules - pass. Test with 2 1GB modules and one 512MB module - fail again. Some things just don't perform up to their advertised spec - sad, but true.

thanks for the response. So to start, i decided to just run with with all dimms filled. I got an error at the end of the first pass/beginning of second. If i remember correct, you cannot really get and info from the error message, right? (ie, which dimm or module it resulted from).

Also, about how long do you think I should run it? When testing ram settings, I generally preferred over night, but now, thinking about all the combos of arrangements, it would like a long time.

Thanks again.

I always assume RAM testing will be a long process. I'd do each run for at least 2 hours. I've had RAM/motherboards that were bad that would Memtest fine for 20 minutes as a quick check, and not develop errors until they were really starting to heat up and they'd fail after 1-2 hours instead.

Post a BIOS template showing all of your current overclocking related settings (all voltages, RAM timings, FSB : DRAM ratio / DRAM frequency, etc.) And I assume you're referring to the GA-965P-S3 that's listed in your sig.? If RAM sub-timings are set to AUTO, post a pic. w/ MemSet open showing the current timings. You might also want to post pic's w/ CPU-Z open to both the Memory and SPD tabs.

redduc900 said:

Post a BIOS template showing all of your current overclocking related settings (all voltages, RAM timings, FSB : DRAM ratio / DRAM frequency, etc.) And I assume you're referring to the GA-965P-S3 that's listed in your sig.? If RAM sub-timings are set to AUTO, post a pic. w/ MemSet open showing the current timings. You might also want to post pic's w/ CPU-Z open to both the Memory and SPD tabs.

Click to expand...

will do tomorrow. yes to the board, is Rev 1 too..

Here is just some more info. I did run the board pretty hard and I understood when I purchased it that it is only a couple of year board. So anyways, at the end of last spring, I started having BSODs and lock ups. I attributed it to too much NB strain and board wear. As a result, I dropped back all settings to stock (FSB, multiplier, vcore, GMCH volt, ram multiplier to auto and timings to 5-5-5-15). I did leave the ram voltage at +.3 though since they run at 2.1.

Trying to remember before the issues, I think I was running:
8x400 FSB, 1.29ish vcore, +.2 GMCH, 1:1 ram and 5-5-5-12 timings. When I was 2x1GB ram, I could run 4-4-4-12, but with 4x1Gb I could not.

So, decreasing all of the settings to stock, excluding the ram voltage, things improved, although there were still some random, intermittent issues. However, with work, I did not have time to delve into the problem.

So here is what I have done so far: running all four dimms at the settings I described as "stock" above, I get error in memtest 86+ at about an hour. Since then, I have only had time to run one module in dimm 0, which was error free up to 4.5hrs.

I am going to test one more individually to verify it is clean, and then run the two in tandem arrays, as suggested.

Ultimately, I think my board is just really on its way out and cannot handle managing four dimms. There were stints of time I did run (for not too long) at 9x400, +.3 GMCH and high vcore.

As well, some other evidence that makes me think board: sometimes it freaks out and doesnt turn on all USB or other ports; such as, I cannot use my mouse and keyboard and have to restart. As well, other times, after changing out ram or something along those lines, the display will not work for a couple of hrs. Also, what prompted to reducing to stock settings, was my board continually for two days locked up when loading bios. Then, shortly after, it just worked fine again. All of this makes me thinking the board is failing.

Thanks guys, will run some more test and upload the other requested info if still needed tomo.


and also, stock ram freq its just set to auto in the bios, so whatever ratio gives me the 800mhz, and rest of the specs are in my sig. Also, to theflyingrat, I am confident it is not the PSU; i have interchanged a few in the past year.

so for the past day or so I have been having the issue described above where the video does not turn on on boot. If I didn't say this above, I believe its an issue with the board not managing ports/peripherals (or however you would term it) properly. Maybe it couldn't react to the quick changes in the amount of ram, idk. When i went to 2x1gb in dual channel, it started the no vid on boot. When it does this no vid on boot, you can see it also is not powering the USBs (ie my mouse and keyboard don't light up).

In the past, this issue normally solved within a few hours and a few cmos resets, however, this time, it did't fix until I downgraded to 1gb of ram. When it finally did a normal boot, an an error stating "CMOS error" popped up.

Sounds like I need a new board?

Thanks guys

what NB do you guys recommend I look at? I don't really want to spend too much money, but I dont want to downgrade performers/options wise. I am considering grabbing a DDR3 board. eh, looking at newegg, the 150$ is a little high... thoughts?


how about this board

That board is okay.....it's P43 not P45, so maybe not, check out this :

http://www.cpu3d.com/review/4924-1/asus-p5q-e-p45-motherboard/introduction.html

Some more boards :Foxconn Black Ops, or any P45 chipset, P5Q-e, P5K deluxe, any of them.

I would reccomend the P5Q-e, but thats only because my one makes my quad fly!!

If I had to guess, I would say it is my ram... I just got the new baord, the giga P45 and it keeps restarting every two seconds when all four modules are in. It boots, however, fine, when two are in.. very strange..

OCZ wouldnt happen to have a lifetime warranty, would they? lol

with the P45 i shouldnt have to increase NB volt to handle 4 modules, right?

I believe most of OCZ's memory has a lifetime warranty of some sort. Contact them - they've been good about replacement parts to me in the past.

Have you gotten to run Memtest+ on it at all yet?

before changing the board, all i got to was running all four, which failed in an hour, running 1, which passed, and running 2 in dual channel, which passed.

I emailed OCZ and opened a ticket; I will see what they say. Thanks. When I gettime, I will run the other two in the mix..

I am tempted to just buy a set of 2x2GB right now though.