Memory failures with Mem torture & 2D/3D Tests

[mbtester]

I'm currently using Version 4.0 and I have much higher incidence of Memory failures with Memory Torture + 2D/3D test than when I have with Full CPU suite + Memory Torture + 2D/3D test. In both cases all tests are set to 100%. This seems to increase as CPU and FSB's increase also.
Chipset is Intel 945G, OS is WinXP SP2. Any ideas that may explain this?

Are there any improvements, new features in the current version BIT 5.2 for DDR2 memory testing to support the newer Chipsets and Dual/Quad Core CPU's? Do you plan to support new tests for Intel AMT technology or Trusted Platform Mgmnt (TPM)

I'm considering updating my site license to the current version. Do you have info on the improvements or new hardware supported from 4.0 to 5.2

Thanks
mbtester

[David (PassMark)]

All the changes between V4 and V5 of the software are documented here,
http://www.passmark.com/products/bit_history.htm

The CPU tests are not very RAM intensive. So if you are only interested in testing the RAM or putting load on the RAM, leave the CPU tests off.

The 2D and 3D tests would add to the load on the bus, so maybe the RAM itself it OK and it is the bus at fault. Or maybe the EMI from the video card when under load effects the RAM. Or maybe the power supply is not so good and when the video card is under load the RAM isn't so stable.

Dual and Quad core CPUs are supported in V5. There were no specific changes for new chipsets, but we support all the current ones (at least the ones that are Windows compatible that is).

Intel AMT is mainly for asset management and remote PC control. I don't know if it is really applicable for a product like BurnInTest.

At the moment there is nothing in BurnInTest for the TPM. We might include something for the TPM in the future. Once people start asking for it and it becomes technically possible. At the moment there are 1000's of pages of specifications for the TPM, but no standard interface code as far as I know.

[mbtester]

Thanks for the info. In this case there is no video card and am using the Intel 945G on board graphics but also have Conroe CPU's. So the intent is to stress it and the memory in doing both Memory and 2D/3D video at the same time. From what you describe, Sounds like either the Chipset device/solder is weak, FSB is noisy &/or need to upgrade to at least Ver 5 to have latest CPU support. How much does BIOS timing play into this? It is Phoenix Ver 6 core. Something is very marginal. My gut says some of this is real, some is false. Need to eliminate some variables as the failures occur anywhere from 4 to 40 hrs at random memory locations. Your thoughts?

[David (PassMark)]

I would start by swapping the RAM and seeing if the problem only occurs with 1 type of RAM or all types of RAM. Then I would see if the problem still happens if only the RAM test is used.

[Ste]

I'm having a seemingly similar problem with the same Intel chipset. See my post at: http://www.passmark.com/forum/showth...=3472#post3472