24 hrs stable in memtest but not in Windows?

[patriot71]

I have been working like no other on overclocking my A64 3200 (CBBID )to get it absolutely stable in windows @ 2.5ghz.

I tested my GSkill PC4400LE for 24hrs in memtest @ 278mhz 2.5-3-3-7 , 2.7v. Not one error. I the boot into windows with 278x7 1:1 for appx 2 ghz stock clock on the cpu so it isnt too far out of spec. With this setting I can only pass SuperPi 8M-32M at 2.5-4-3-7 in windows at that setting.

I finally got the cpu 32M SPi stable at 2.5 ghz 278 x 9 1:1 but I had to release the timings to 2.5-4-4-7. Chipset voltage @ 1.6v & LDT voltage @ 1.3v. Is this an example of the infamous memory controller coming into play or could their be another factor keeping me from 2.5-3-3-7 in windows?

DFI SLI DR w/ 3.10 bios
2 x 512 GSkill PC4400LE
OCZ Powerstream 520w
A64 3200+ CBBID

[Jackass]

CBBID memory controller strikes again

[craig588]

Are you running test 5?

[Jackass]

test 5 is good for finding memory faults when oerclocking, but it sucks when finding yoru mem controller's windows max, for this I would more recommend using the windows memtest and other windows based stability apps

[Aguanteboca]

Thats why I dont waste my time with memtest anymore. I just use it to get a very basic pre stability test for like 20 runs of test 5.

[patriot71]

Are you running test 5?

I looped test 5 for about 7 hours with no errors and then I came back later and re-adjusted the test to run all tests overnight. No errors.

[saaya]

hey, this thread doesnt really belong in the xtreme overclocking section wich is for xtreme overclocking topics only, i moved your thread to the xtreme bandwidth section wich is for all memory related topics.
please post in the acccurate section the next time, if you dont know what section to post in the please contact any of us mods and we will help you

oh, and please use the search button and try to find an answer to your question before posting a new thread, there are many threads about this topic and there are some nice guides that can help you

[S0nic]

looks like a memory controller issue, blame ur CBBID winnie..

[ben805]

like other said, I have had the same experience in the past as well, my previous XP-M was memtest stable for hours, but crashed prime95 and BSOD on windows, sometime it's the other way around and memtest would get thousands of error but prime95, 3dmark and windows were fine. I don't really trust memtest anymore, as long as your windows doesn't crash on 3d bench, superpi 32m and prime95, then your rig is stable.

[NeoGeo333]

Like the two above, its probably the mem controller i got a cbbid and it woudnt do squat. now i got a 3500 clawhammer and im running 2.65ghz with mem at 295 x 9 at 1.68v. As soon i hit it to 2.7ghz it would just hang, maybe its time to get me a water setup.