Video card causing Memtest86+ memory errors?

M

Maniac78

n00b
Joined
Mar 6, 2012
Messages
8
Is this possible? I posted on Tom's and people seem to think it's the video card's problem but I'm not 100% convinced so would like your guys council.

Here are all the details, sorry for being so long-winded but I tried to explain everything as best and as completely as I could.

-------

I have the following system specs:

  • Windows 7 x64 w/ all the latest updates and drivers
  • Intel i7-4771
  • MSI H87-43G
  • 1x XFX R9-270X-CDFC Video Card
  • 2x4G G.Skill F3-12800CL8-4GBXM DDR3 configured with XMP*
  • Corsair CMPSU-650TX
  • 1x Intel 840-EVO 256GB
  • 1x 1TB SATA Western Digital Black
  • 2x 500GB SATA Western Digital Black
  • 3 case fans
  • 1x DVD-RW Drive

*NOTE: I checked the XMP detected settings and they are correct.

Background:
I got a few BSOD referencing memory errors (can't remember the exact error but it was definitely memory related). I originally had 8 GB of G.Skill Ares memory and Memtest86+ indicated a lot of errors so I replaced that for the above listed memory. When I got the new memory I ran Memtest86+ 5.01 right off to make sure all was good but I got lots of errors immediately however Windows now seems fine with no blue screens, etc.

I wanted to narrow down the Memtest86+ errors so I took the video card out and unplugged the 3 SATA hard drives. I also took out one stick of memory and then re-ran Memtest86+ and it was fine, I then took out that stick of memory and tried the other and it was fine as well.

To narrow it down further I put the video card back in but left the drives unplugged. When Memtest86+ got to test 7 I got a lot of errors. So I again removed the video card, used the built-in Intel display, plugged in the drives, and got no errors at all. Then I ran Memtest86+ one last time with all the above listed component and hooked everything up and ran Memtest86+ and again got lots of memory errors.

One other odd thing I noticed is that when you first run Memtest86+ it asks you to press F2 if you want to use multiple cores. If I do this with the video card installed it will freeze if I try to access the configuration menu but if the video card is removed, and I just use the built-in Intel, then I can run with multiple cores and access the configuration menu just fine.

The AMD Reference Design Radeon R9 270X graphics card AMD specifies a minimum of a 500 Watt or greater system power supply. The power supply should also have a maximum combined +12 Volt continuous current rating of 30 Amps or greater and have at least two 6-pin PCI Express supplementary power connectors. My PSU is rated for 650 Watts with 52A on the +12 volt (if I'm reading things right): http://www.corsair.com/en-us/tx650w
Corsair TX650W Specifications
AC INPUT 90-264V ~ 9A 50/60Hz
DC OUTPUT +3.3V +5V +12V -12V +5Vsb
MAX LOAD 24A 30A 52A 0.8A 3A
MAX COMBINED WATTAGE 170W 624W 9.6W 15W
TOTAL POWER: 650W

So it seems I should be fine, right? Or are my other components pushing me over the limit? Could it be something else all together?

Thank you very much for any help you can give, I'd really appreciate it!
 
Could be a faulty card, or PSU. Specs wise the PSU should be overkill.
 
Have you tried toggling the setting for USB Legacy Support in your motherboard's BIOS? It may go by similar names like Legacy USB.

Have you tried using Memtest86 instead? Memtest86+ is a fork from when Memtest86 didn't get updated for a long time, but Memtest86 is now being maintained again. They look similar but often return quite different results.
 
If you really think it's video then pull the video card and run off of the integrated for a bit, see if anything changes.

Having said that I rather doubt a video card can cause problems with Memtest though.
 
Thanks for the replies.

@FnordMan: Yes, I tried with the integrated and all was fine. Put the video card back in and got memory errors.
I also agree that it's odd for a video card to affect memtest.

@evilsofa: USB legacy is enabled in the BIOS. I tried Memtest86 as well as Memtest86+ and I even tried the Windows 7 memory tester that will run when you reboot before Windows loads. Here's what I found:

Windows 7: No errors at all ... took hours to make a single pass but no errors. Didn't have time to try more than one pass.

Memtest86: Finds an error always in the same basic place: 3355.xx MB. Usually finds it between the first and fourth tests but never after that. If I restart the tests from the configuration menu it won't find the errors again it also won't find the errors on subsequent passes, only that initial pass. If I totally restart the computer and start the test over again then it's the same thing, first pass shows errors, subsequent passes don't.

Memtest86+: Very similar to Memtest86. Finds error around the same place always before the 4th test but not after, subsequent passes find nothing. If I enable SMT support (F2) and then try to configure options Memtest86+ will freeze and require a hard reboot.


Keep in mind that Windows it's self seems stable and after replacing my initial memory with the new batch I have not had any crashes or BSODs. I only see these weird errors in Memtest86/86+ and the only reason I even ran the tests was to make sure I got a good set of memory seeing as the first set was bad.

It's hard for me to understand how a video card could affect a memory test, it just doesn't seem to make sense but I can't say I know that much about all this. Could the motherboard maybe have issues? Should I see if I can get NewEgg to replace the MB and video card under the 30 day exchange warranty just to be safe or is that not worth it?
 
So I did some more testing: I dug up an old ATI HD 4870 and ran both memtest86 and 86+ and neither reported any errors after one pass. I also ran one memtest86+ pass with my new XFX R9 270x and just one stick of memory and got the same errors at 3255.xx MB then again with just the other stick and got the same... errors at 3355.xx MB.

An interesting note is that it took about 30 minutes to complete one pass with the HD 4870 but only 18 minutes with the R9 270x.

So all this leads me to think that maybe it is a problem with the video card though I just don't see how that's possible. The only other things I can think of are that the motherboard is bad or incompatible with the video card somehow or that the PSU is maybe getting old or going bad and not quite supplying a stable current. However if it is the PSU then it seems the errors would be much more random and not always at the same spot.

I'm considering switching out the video card while it's still within the 30 day return window and see if that helps... Unless someone has another brilliant idea. :) Do you think I should replace the motherboard as well to be safe or just one step at a time? I only have 15 more days before the 30 are up and I have to RMA which I'd love to avoid if possible.

Thanks for the advice!
 
Good news! I found that MSI just recently released a BIOS update and after installing it I was able to run 2 passes of Memtest86+ without any errors. Thanks to all for the help and tips!
 
You must log in or register to reply here.
Back
Top