Memory fail

didgemaster

didgemaster

didgemaster

Associate
Joined
11 Aug 2003
Posts
1,233
Location
Taunton, Somerset
Memory seems to fail.

I have been having problems : I have 6GB of OCZ obsidian that fails under certain tests. Under Windows XP (32bit) it runs stable - I assume its because it only sees the first 3.4GB of RAM. Under windows 7 64bit it fails.

So I have done some more testing and it seems to have come down to one of the strips of memory. With memtest x86 it doesn't seem to fail but when I run prime95 under windows 7 64bit - build 7100 it fails. I am going to try it under windows XP to see if it fails.

Does anybody know of other memory testing software that works in and/or independently of Windows.
 
Which tests does it fail?

I have "issues" with my memory, so it fails Prime95 and OCCT in Win7, but passed 5 hours of memtest86+ (from a bootable cd)...

Currently troubleshooting but it may be a case of voltage or timings just running slightly out of spec.

What is your memory rated at in terms of speed, latency and voltage, and what are you running it at in the BIOS?
 
Rated:
Speed = 1600MHz
Latency = 9-9-9-24
Voltage = 1.65V

I have tried running BIOS settings with:
1600MHz, 9-9-9-24, 1.64V
left to full auto with 1.64V
left to full auto

all fail under windows 7 64bit build 7100, all pass memtest x86, currently I am testing under XP - I'll let you know how it goes
 
OCZ will tell you to try a few things:

Command Rate = 2T
tRFC = 88

Uncore (QPI/VTT) voltage to 1.35V, but this depends on the motherboard/CPU. I have a Lynnfield Core i5 and intel specs max out at 1.21V. So I have left this on Auto, but may notch it up to 1.15-1.20 to keep it within 0.5V of the memory (at 1.64V, seeing as my mother has 1.64 or 1.66).

The thing that seems to bring down my stability is the AUTO base clock (BCLK). Check my other posts, but seemingly setting the base clock to manual (133) improves the mem stability (auto give BCLK 136.7). I'm still in the middle of testing things, but so far so good.
 
Well the strip in question failed in windows XP so I thought I would leave it on a memtest run while I was in work. 10 hours later and 27 passes not 1 error however this is on auto.

QPI 4.8GT/s
BLK 134MHz
Timings 7-7-7-19
speed 539MHz (1079)

Will try your suggestions to see if it makes any difference
 
kidscruff said:
OCZ will tell you to try a few things:

Command Rate = 2T
tRFC = 88

Uncore (QPI/VTT) voltage to 1.35V, but this depends on the motherboard/CPU. I have a Lynnfield Core i5 and intel specs max out at 1.21V. So I have left this on Auto, but may notch it up to 1.15-1.20 to keep it within 0.5V of the memory (at 1.64V, seeing as my mother has 1.64 or 1.66).

The thing that seems to bring down my stability is the AUTO base clock (BCLK). Check my other posts, but seemingly setting the base clock to manual (133) improves the mem stability (auto give BCLK 136.7). I'm still in the middle of testing things, but so far so good.
Click to expand...

Command Rate = 2T
tRFC = 88
Uncore (QPI/VTT) voltage to 1.35V
BCLK =133

still failing

I am going to try memtest again with these settings
 
Stable under memtest so I'm now using the good to strips to see how stable they are. so far I've run prime95 for 2 hours - I'll leave it going overnight and update tomorrow

I have a post on the OCZ forum to see if they can help
Link
 
You must log in or register to reply here.
Back
Top Bottom