Confusing Test Failures in 2GB D875PBZLK

[kengregg]

I need some help in trying to diagnose whether I have one or more bad modules in my newly-configured 2GB system. Any leads would be appreciated. Here are the details...

 

I have built my 3rd system based on the Intel D875PBZLK (newer Prescott version) motherboard with an Intel 3E 3GHz Prescott CPU, and have run into a pattern of failing memory tests that I can't figure out. The memory consists of 2 brand new sets of TWINX1024-3200C2PT modules (with the correct manual BIOS settings for this board: freq = 400, CPC override = disabled, 2-3-3-6, volts = 2.7). The matched pairs are distributed correctly between DIMM channels on the motherboard. The motherboard has the latest and greatest P33 BIOS. The test is Memtest86+ v 1.40. All things are equal to the two former systems I've successfully built, other than the fact that all the hardware is new.

 

The errors occur only when I run will all 2GB in place. The errors occur only in tests 5 and 7 (random number sequence)...all other tests are okay. The errors do not occur at all when I test each module individually (512MB) overnight, or when I test each pair without the other pair (1MB) overnight. But when I put all 2GB together, both tests 5 and 7 have failures on every interation of the suite.

 

The failing addresses vary, showing up at as low as 182MB, and as high as 1766MB, so I can't pin down whether the problem is in the lower bank or upper bank. The failing bit is always the same...00000004. The BIOS and the test report that it's running in dual-channel mode. At the moment, it's been running 2:23:21, it's on pass 4, and it's reported 40 errors so far. Most of the failures are occurring duing test 7, but both tests 5 and 7 contain failures. In both cases, the addresses are all over the map, but again the failing bit position is always the same.

 

Any ideas on what I can do to diagnose what's causing these errors? Is this more likely a motherboard problem, because the individual modules and pairs test out okay? Any ideas are welcome. Thanks in advance for your help.

 

- Ken

[RAM GUY]

With 4 modules I would suggest you set the timings to By SPD or Cass 3-3-3-8 and test the system with http://www.memtest.org. If you get errors please set the Dim Voltage to 2.775 Volts and see if that solves the problem. If not please test the modules one at a time with http://www.memtest.org at the tested settings Cass 2-3-3-6 and lets make sure one is not failing.

[kengregg]

Question: When using 3-3-3-8, should CPC Override still be disabled?

 

I will update to the new version of Memtest86+ v1.51, run the tests as you suggested, and report the results.

 

Note that I already did test each module individually using 2-3-3-6, as well as testing the individual pairs at that setting. When only one or two modules are present in the system, all the tests pass for hours on end. Only having both pairs present leads to the test failures described. I'll retest with v1.51 of Memtest86+ just to make sure, though.

 

Since this is the third identical system I'm building, with the same motherboard rev, BIOS version, 2GB memory size using two TWINX1024-3200C2PT sets, etc. in all the systems, I'm curious about why the 2-3-3-6 setting would fail in this system, but works perfectly in on the other two systems.

[kengregg]

OK, I left CPC Override disabled, changed the settings to 3-3-3-8, and used the 1.51 version of the test. No errors reported in 3 test passes. I did not alter the voltage setting.

 

Questions:

 

(a) Should I leave CPC Override disabled with these settings?

 

(b) Does using 3-3-3-8 mean that this new machine will have slower performance than the other two identical machines that use 2-3-3-6 (which also have four modules of TWINX1024-3200C2PT in each system)?

 

© Is there an explanation for why identical rev motherboards with all other components the same would require that this new machine use 3-3-3-8, while the other machines work great with 2-3-3-6? I'm very curious about this. Does this imply that there is some defect in the new motherboard, or in the new memory modules?

 

Thanks for any info.

[RAM GUY]

If you have another system I would suggest testing these modules in that system and see if the errors follow the modules. If so please test them one set at a time.

[lektrix]

Hi, I have almost exactly the same problem on my new system. Specs are as follows

AMD64 3500+

MSI kn8 NEO2 PLATINUM mobo

2 pairs of TWINX1048-3200C2PT (2Gb)

xfx 6800ultra

550w PSU

 

In memtest 3.2, I have run tests on each stick on its own with settings at 2-3-3-6 with no fail.

I have run tests on two sticks together with same settings with no fail. both sets, and I mixed them up and tried again with no problem.

But when I put all 2gig in I get problems, round the 40% test mark at 98% on test 5, I get lots of errors ranging from as high as 1600meg all the way to 0.0meg. 268 in all.

 

I've tried all different voltages and all settings to default at 3-3-3-8 and still get the same errors.

Could this be the motherboard?

Unfortunatly I have no other machine to try it in.

The fact that all dimms pass the memtest with flying colours up to 1gig leads me to believe its not the ram.

What would you recommend I try/do?

 

Thanks in advance!

 

Max

:sunglasse

[RAM GUY]

Can you tell me the make and model of PSU? And I would maybe try one that is ATX 2.0 Compliant. And I would set the Dim Voltage to 2.8 Volts at 2T Command Rate and set the timings to Cass 3-3-3-8.