memtest sort of freezes

[rnjeezy]

Running memtest after setting the memory timing as specified and somewhere during the first pass memtest appears to be halted but the '+' sign next to the memtest title in the upper left continues to blink.. Is there some significance? (If it was a complete halt wouldn't the plus sign itself stop blinking?) Is this type of memtest freeze indicative of something? (There don't seem to be any memory errors, but it does freeze before the first pass is complete, and this (memtest) has been attempted a few times)

Edit: When I say memtest halts, it basically crashes during a memory test, seemingly indicating the memory/settings are unstable. But the + sign keeps blinking, so i guess it's not a total hardware crash.

Edit2: Other than this, someone on OCForums mentioned that it may have to do with having legacy usb support enabled..

 

[rnjeezy]

Looks like disabling legacy usb support in the mb bios resolved the issue. this was on an i7 build, btw.

 

[LiuKangBakinPie]

did you run memtest inside windows?

 

[VirtualLarry]

The blinking text is handled by the video card, autonomously from the CPU. If the CPU crashes and freezes, the text will still blink, unless the video card's processor crashes.

 

[rnjeezy]

The blinking text is handled by the video card, autonomously from the CPU. If the CPU crashes and freezes, the text will still blink, unless the video card's processor crashes.

Thanks,I was thinking of this possibility as well. In any case, I disabled legacy USB support, it didn't crash on the test that I did. I did only one test, and it finished 3 passes, so i guess it could have still been a crash, and it had just randomly happened to work that time. Or that those other three stoppages during memtest prior to this were actually due to instability. I mean, it hadn't crashed before when I tested it at ddr1066, although I did get numerous errors at ddr3 1066 in memtest until I changed the settings as indicated below, which are the rated settings forddr3 1600.

did you run memtest inside windows?

Memtest86+ 4.20 was run from the boot menu after POSTing. I would like to know of a good app for testing stability in windows with extensive memory usage, though. Anyway, it turned out Windows was .. unstable. First time I went to windows and tried to play a video file the computer crashed. Second time I wasn't able to open files on a hard drive. Third time it just froze while I was transferring files to it over the network. I think this has to do with the memory issue. I went back into the bios and lowered the speed from ddr3-1600 to ddr3-1066, which was what I had been using before. ddr3 1600 is the rated speed of the memory. I'm optimistically thinking perhaps I'm getting the settings wrong, that nothing is seriously broken with regards to my parts. I have everything on default except for 8-8-8-24 , qpi voltage to 1.35v, memory voltage to 1.66V. I'm using an asus p6t deluxe v2 and a OCZ3G1600LV6GK

Edit: scanning the ocz forums, and in some thread, one of the staff mentions tRFC - 88, Command Rate - 2N . I'm pretty sure it was tRFC - 88 by default, and 1N. I guess i'll try it on 2N. Not sure why they don't put this in an easily accessible location..

 

[VirtualLarry]

Just so you know, OCZ RAM is crap. So bad, in fact, that OCZ left the RAM business.

 

[rnjeezy]

I figured, but this ram was purchased a long time ago. I actually got an RMA back due to one of the sticks from the original set dying, (the issues I'm having are with this 'new' set). After having looked at amazon / newegg, the product ratings, to be taken with salt, for this set are not reassuring, but not abysmal either.

I just ran the prime95 test which uses alot of memory, and it gave me a fatal error after 4 min, the computer didn't crash though. Not sure what kind of action I should take at this point, besides underclocking it again for stability's sake.

 

[LiuKangBakinPie]

Thanks,I was thinking of this possibility as well. In any case, I disabled legacy USB support, it didn't crash on the test that I did. I did only one test, and it finished 3 passes, so i guess it could have still been a crash, and it had just randomly happened to work that time. Or that those other three stoppages during memtest prior to this were actually due to instability. I mean, it hadn't crashed before when I tested it at ddr1066, although I did get numerous errors at ddr3 1066 in memtest until I changed the settings as indicated below, which are the rated settings forddr3 1600. Memtest86+ 4.20 was run from the boot menu after POSTing. I would like to know of a good app for testing stability in windows with extensive memory usage, though. Anyway, it turned out Windows was .. unstable. First time I went to windows and tried to play a video file the computer crashed. Second time I wasn't able to open files on a hard drive. Third time it just froze while I was transferring files to it over the network. I think this has to do with the memory issue. I went back into the bios and lowered the speed from ddr3-1600 to ddr3-1066, which was what I had been using before. ddr3 1600 is the rated speed of the memory. I'm optimistically thinking perhaps I'm getting the settings wrong, that nothing is seriously broken with regards to my parts. I have everything on default except for 8-8-8-24 , qpi voltage to 1.35v, memory voltage to 1.66V. I'm using an asus p6t deluxe v2 and a OCZ3G1600LV6GK

Edit: scanning the ocz forums, and in some thread, one of the staff mentions tRFC - 88, Command Rate - 2N . I'm pretty sure it was tRFC - 88 by default, and 1N. I guess i'll try it on 2N. Not sure why they don't put this in an easily accessible location..

OCZ ouch! They left the ram business and for good reason. Asus boards are very picky with the ram you put in them. Did you take the ram of auto and set it to 1.5v manually?

 

[rnjeezy]

The ram is, (was), at the settings according to OCZ. That is, 8-8-8-24 / 88 (tRFC i believe) / 2N (command rate) / dram voltage 1.66V / qpi voltage 1.35 / DDR3-1600 MHz. However, the system is unstable at these settings. I ran Prime95 and got errorswithin 5 min.

After running prime95 for ~15Hrs, and getting no errors, I can say the system appears stable at 8-8-8-24 / default [Auto] (tRFC i believe) / default [Auto] (I believe 1N command rate) / dram voltage 1.66V / QPI voltage 1.35V / DDR3-1066 MHz. I will probably leave it at this rate for now to test for continued stability, and potentially nudge the settings, increasing them performancewise, until I reach a point where I'm having stability issues again.

I would post this on the ocz support forums, but have a feeling it's not going to be useful.