• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

RAM Corruption

H

haxxorboi

Limp Gawd
Joined
Aug 21, 2007
Messages
159
Specs:
Mobo: Gigabyte P35C-DS3R (latest bios)
CPU: Intel 6750 w/ TR 120-extreme
RAM: OCZ 8500-SLI
PSU: Seasonic Energy-Plus SS 650 watt

I've had my PC overclocked to 3.6Ghz @1.4v with the RAM running at 1125Mhz @ +.3v (which should be 2.1, but for some reason always ran at 2.24v) I orthos blend tested it for over 10hr and it was completely stable. Yesterday while playing CoD4 it BSODed and I checked all voltages and temps, all which were fine.

So this morning I decided to run Mem-Test+ and within about 5 seconds I had over 25 errors. The output block that was erroring was 0007feedf18-2046.8MB-Good(00000000)-Bad(00000001).

I'm guessing that this means that the RAM went to hell... however I don't exactly understand why this might have been. The RAM specs are for 2.3v stock. Also, if anyone can enlighten me as to why the mobo sets the vDimm higher than it should be... I've seen the same issue from other owners of the mobo, too...

Thanks in advance and Happy New Year!
 
Yep, I did. I set everything back to normal, even reset the CMOS, and it still returns the same results. Heck, I even underclocked the RAM down to 800mhz and it still errors... I also tried to run the RAM at +.4 vDimm (ran at about 2.34v) and it still errored...
 
good news, apparently this is not a localized issue... apparently this overvoltage issue has been occurring since September and Gigabyte hasn't really done anything to fix this issue... great support...

Also, as a note to Gigabyte P35 owners, if you've upgraded to F9 bios you really want to read this: http://forums.tweaktown.com/showpost.php?p=246311&postcount=9 Basically, go back to F8...

This is becoming a very bad beginning to my 2008, but perhaps my RAM isn't fried... I'll try to update after I back-flash and retest...
 
Um, crap...I just set up my build, didn't really pay attention to the bios version (figured I'd tweak everything to my liking first, and would only update if I wasn't satisfied). I just started tweaking mem settings (as detailed in my sig...so far so good w/ blend in Prime95...going to let it run overnight). Is the issue the board giving the ram more juice than specified?
 
Basically, the vCore is getting undervolted and the vDimm is getting overvolted. The core isn't so much of an issue, as long as you're not volting to like 1.6v on stock or something... however when you set the board to do +.3 (2.1v) for some 2.1v dimms and it gives you 2.24v... well there's a good chance you can fry your new shiny RAM...

Luckily, my OCZ is rated with lifetime warranty to 2.35volts and I'm very happy of this fact... probably saved me...

Oh, and I just reflashed down to BIOS F2, which apparently people are having the most stability and overclockability (though that info was from about a month or so ago, but straight from OCZ forums.) Had it run Memtest+ for 2 hrs from UBCD, not an error in sight!! Just bumped it to 3.68Ghz and around 1104mhz RAM (set to +.3, should still be safe there... probably should check that) and both Memtest and Prime are going great...
 
Oh, vcore's affected, too? Maybe that's why cpuid says it's 1.44 even though it's set at 1.46 in the bios...cpuid shows my vdimm at 2.1 (currently set it at +0.3) and my ram' s vdimm is rated at 2.2 +/- 0.05%...so I figured I was ok. Everything did well overnight at least, though when I powered down today to fiddle with a fan and powered up, windows said it 'recovered from a serious error' (wtf?) so maybe i'll take it down a notch...

From your results and from that forum, it seems that these memory errors are false positives from some bug in the bios rather than memory getting fried, though.

Oh, and I think I have the F8 bios...
 
Yes, the F9 bios creates false positives, however the fact that the RAM is putting the voltage higher than it should is shorting out a lot of people's RAM. People don't realize how much is going into their RAM and think +.4 is going to do 2.2 when it really does about 2.35v... If your RAM is rated for 2.2 and you run 2.35 into then you're bound to burn them out REAL fast unless you have some nice cooling on them...
 
Well, I now set it at +0.25, and cpuid reports my vdimm as 2.05 (well within my ram's specs)...I'm not even sure I should bother lowering timings on this memory...I was only able to get down to 4-4-5-15 from 5-5-5-15 anyways...

so how exactly did you find out your ram was running at 2.24 instead of 2.1?
 
I saw it in BIOS Health, then checked Everest, then SpeedFan, then checked with a Multimeter on the board (it was a little different, but still right around 2.24.) I kinda freaked when I saw the memory errors and then started looking online, which is when I discovered the errors...

Currently I'm kinda annoyed, because my OCZ ram is spec'd at 2.3v, and I have my proc at 3.8Ghz stable with the ram running at 950mhz (1:2), but because I can't get the voltage up beyond 2.3 without going over 3.5v I can't run it at 1140mhz (1:2.4) or 1187.5 (1:2.5) which others have been able to get this ram to... kinda annoying really... guess I'll wait for F10 and see if they get their crap in order... or maybe that tech will actually respond from the R/D dept at Gigabyte...
 
Well, with +0.20 both cpuid and my bios health status both say my vdimm is ~2.00...I do have F8 btw
 
I went down to +0.15 (showing as 1.95-1.98 in bios health + hwmonitor)...things were stable 24/7...

Does your bios/monitoring software still tell you exaggerated voltages?
 
You must log in or register to reply here.
Back
Top