BSoD, x850xt, and new memory

[Gewesen]

I recently bought an ATI Radeon x850xt and a new 512 stick of memory from Crucial. I updated the BIOS, reset the system back to its original configuration (used the Falcon restore disks), applied updates to the OS, uninstalled drivers for old video card, switched cards, installed new video drivers (from the CD).

At this point, everything ran nicely. After about a week or a little more, I received my new memory. I've triple checked to ensure it is correct for my ASUS A7V133 mobo. I got BSoD, and ran Memtest86. I get bad results with both chips in, but good results for each of the single chips. When I reverse positions of memory sticks, I still get BSoD, but maybe not quite as often.

Crucial suggests updating video drivers (and I haven't checked yet to see if I have the most current ones or not)....but I'm wondering if anyone has a suggestion on how to identify what the problem really is? Memory stick bad? Memory slot bad? Are the two memory sticks incompatible somehow? Video card or CPU too hot?

Argh.

Any help or useful opinion would be much appreciated.

 

[Powermoloch]

Try test one of the sticks at a time.....that's my recommendation.

 

[Gewesen]

Yep, I did that. Each stick checks out fine by itself. When used together, I got several errors...and I guess I need to do that again to confirm if Memtest is working correctly. For example, the Memtest docs say:

"There are some systems that cause Memtest86 to be confused about the size of memory and it will try to test non-existent memory. This will cause a large number of consecutive addresses to be reported as bad and generally there will be many bits in error."

The first time I test both modules, with the new module in the second slot, I got multiple errors on the first test...within a half second. The next step was testing was both modules, after reversing their sequence in the slots -- that showed no error. Then I tested each individually...both showed no error. I ran only the individual tests for prolonged periods.

So far, the initial BSoD in each session occurs during play of World of Warcraft. Thereafter, the BSoD has occured when launching WoW again or when shutting down...but not always.

I'll have to run some more mem tests tonight, I guess, after I investigate updating the video drivers (as Crucial suggested).

Any other ideas are most welcome...

Thanks...

 

[Fern]

Any other ideas are most welcome...

Back off on the rams' timings (loosen them up a little) and see what happens.

Fern

 

[moonboy403]

1) try loosening timings like Fern said

2) bump up the voltage a bit...maybe +.1 or .2v

what's the voltage now?

 

[DetroitSportsFan]

Originally posted by: moonboy403
1) try loosening timings like Fern said

2) bump up the voltage a bit...maybe +.1 or .2v

what's the voltage now?

At bare minimum, check and make sure your ram is running at the voltage its manufacturer states it should be running at. For example, don't try to run high performance ram at value ram voltage (2.5volts) .... Performance ram requires more juice.

 

[Matthias99]

Crucial suggests updating video drivers (and I haven't checked yet to see if I have the most current ones or not)....but I'm wondering if anyone has a suggestion on how to identify what the problem really is? Memory stick bad? Memory slot bad? Are the two memory sticks incompatible somehow? Video card or CPU too hot?

If you're failing memtest, don't worry about anything else until you've fixed that. The system will NEVER be stable as long as you are having memory errors.

If it works with either stick in single-channel, but fails in dual-channel, it could be that the memory controller is being slightly flaky in dual-channel mode. Try lowering timings and/or underclocking the RAM, as recommended.

 

[Gewesen]

I'm certainly a novice when it comes to performing the actions you all have proposed. (I'll look online for more information on that and ask for help/opinions if I need to.) But another thing occurs to me -- power. Could this at all be caused by my power supply? It is only 365w, and the video card alone requires 350. I'm wondering if the additional memory might bump the power consumption enough to cause these problems. The last one read:

STOP: 0x000000D1 (0x00000000, 0x00000002, 0x00000000, 0xE1CB9CB5)
DRIVER_IRQL_NOT_LESS_OR_EQUAL
Beginning dump of physical memory.

So here's my alleged plan:
* Test for heat (I did that informally by taking off side of case - ran into mem error running Wow, then BSoD at system shutdown)
* Test with only new stick in system (it was stable with only 1 old stick of RAM - I'll do that tonight)
* Adjust timings or underclock RAM (need to do homework on that)
* If all else fails, maybe a new power supply (say, 450 or 500w?)

Fern, DetroitSportsFan, and Matthias99: I'll do my homework on timing/underclocking and post again if I need to.
moonboy403: I don't know what current voltage is, but I'll try to find out how to find out.

Thanks to all for your help!

 

[Matthias99]

But another thing occurs to me -- power. Could this at all be caused by my power supply? It is only 365w, and the video card alone requires 350. I'm wondering if the additional memory might bump the power consumption enough to cause these problems.

It's unlikely that adding the RAM is somehow pushing you over the edge -- the RAM doesn't even run on the same voltage rail as the video card, and it uses very little power relative to everything else.

And that card does NOT use 350W of power; they tell people to have a "350W" PSU for it because many people buy extremely cheap and crappy PSUs that don't live up to their wattage ratings (or bought systems that were built with such PSUs to cut costs).

That said, you cannot rule out a bad or insufficient PSU causing your problems. You haven't mentioned what kind of PSU it is, leading me to believe it is probably some off-brand piece of crap. If you want to rule out insufficient power, you could try switching in another video card if you have one, and unplugging everything else (you only need the CPU, RAM, and video card to run memtest).

The last one read:

STOP: 0x000000D1 (0x00000000, 0x00000002, 0x00000000, 0xE1CB9CB5)
DRIVER_IRQL_NOT_LESS_OR_EQUAL
Beginning dump of physical memory.

That is an extremely common error if your memory is being flaky.

If you're failing memtest at all, your RAM is not working correctly. Testing anything else while you are still having failures with memtest is pointless.

 

[Gewesen]

I'm not at my system, so I don't know the PSU brand, but I am sure it isn't crap. The system is an older Falcon Mach V, so all the components should be pretty darn good (for 2 1/2 years ago, anyway).

Are all memory failures due to bad memory? For example, using Memtest86, I had memory failures only once when using both sticks. (As I posted earlier, the kind of errors encountered are similar to those Memtest docs say are due to Memtest problems detecting certain configurations.) I ran 3 other tests with both modules installed and got no errors. Memtest showed no errors for either RAM module, when the modules were tested individually. That led me to think that it wasn't bad memory but either 1) bad second slot, 2) possible incompatabilities between the two RAM modules, or 3) some other unknknown factor related to having 2 modules installed instead of 1.

Crucial support suggests installing new video drivers (I haven't had time to check the installed version, but the last modified date for the applicable drivers at the ATI site is only a few days ago, so I'll try that). Falcon support wondered if the PSU might be the contributing factor. I'm going to try running with only the new RAM module. If I encounter no problems, I'll assume my problems have something to do with using the two sticks of RAM...either the second slot, incompatabilities, the PSU, or something else (sigh). As I've admitted, I'm a novice, so I don't know what else might be the problem.

After I update drivers and run with one stick, I'll move on to lowering timings or underclocking the RAM, as suggested in posts above. I'll probably run some more Memtests with both modules installed, just to see if I can reproduce the errors I first saw.

And I thought this would be a relatively simple last gasp upgrade.

Thanks, Matthias.

 

[Matthias99]

Originally posted by: Gewesen
Are all memory failures due to bad memory? For example, using Memtest86, I had memory failures only once when using both sticks. (As I posted earlier, the kind of errors encountered are similar to those Memtest docs say are due to Memtest problems detecting certain configurations.) I ran 3 other tests with both modules installed and got no errors. Memtest showed no errors for either RAM module, when the modules were tested individually. That led me to think that it wasn't bad memory but either 1) bad second slot, 2) possible incompatabilities between the two RAM modules, or 3) some other unknknown factor related to having 2 modules installed instead of 1.

Any of the possibilities you suggested could be the cause, except that '2' is pretty much nonexistent with newer motherboards.

Crucial support suggests installing new video drivers (I haven't had time to check the installed version, but the last modified date for the applicable drivers at the ATI site is only a few days ago, so I'll try that).

This is stupid, because you are having memory failures in memtest86 without even running Windows. Your video drivers (or lack thereof) will have no impact whatsoever on these tests.

Falcon support wondered if the PSU might be the contributing factor.

Possible, but I find it unlikely that one stick of RAM would push your PSU into instability. If it's a good brand PSU, and you don't have issues with just one DIMM installed, I would be much more inclined to blame the motherboard than the PSU.

I'm going to try running with only the new RAM module. If I encounter no problems, I'll assume my problems have something to do with using the two sticks of RAM...either the second slot, incompatabilities, the PSU, or something else (sigh). As I've admitted, I'm a novice, so I don't know what else might be the problem.

That sounds like a good plan. If you are sure the new memory is OK, then the problem must be either an issue with running two DIMMs at once, or a problem elsewhere in the system.

 

[Fern]

Originally posted by: Gewesen
I'm certainly a novice when it comes to performing the actions you all have proposed. (I'll look online for more information on that and ask for help/opinions if I need to.) But another thing occurs to me -- power. Could this at all be caused by my power supply? It is only 365w, and the video card alone requires 350. I'm wondering if the additional memory might bump the power consumption enough to cause these problems. The last one read:

STOP: 0x000000D1 (0x00000000, 0x00000002, 0x00000000, 0xE1CB9CB5)
DRIVER_IRQL_NOT_LESS_OR_EQUAL
Beginning dump of physical memory.

So here's my alleged plan:
* Test for heat (I did that informally by taking off side of case - ran into mem error running Wow, then BSoD at system shutdown)
* Test with only new stick in system (it was stable with only 1 old stick of RAM - I'll do that tonight)
* Adjust timings or underclock RAM (need to do homework on that)
* If all else fails, maybe a new power supply (say, 450 or 500w?)

Fern, DetroitSportsFan, and Matthias99: I'll do my homework on timing/underclocking and post again if I need to.
moonboy403: I don't know what current voltage is, but I'll try to find out how to find out.

Thanks to all for your help!

I suggest your first step should be going into your BIOS. Check your mobo manual, the BIOS options will be discussed there.

In there will be an option to adjust the voltage to your ram. If not already, try 2.7 volts. Often BIOS will set it automatically to 2.5v

Also, you will likey find an easy way to under-clock your ram timings. There ia likely a choice for "most stable" (as opposed to "turbo" for slight overclocking).

There is likely also a place to set each timing individually (ras-to-cas etc). But by choosing "most stable" (or whatever similar language it uses) they will all be automatically loosened up. That is much easier than fooling around with each of the 4 values.

If you don't have your manual handy, you can prolly google for an online version.

If your still having probs tomorrow I'll try to help. I'll check an online manual and advise. I've gotta run or would help more now.

Fern

 

[Gewesen]

Hey, Fern:

Thanks very much for the help.
Last night I called Falcon support and they walked me through some BIOS changes. Unfortunately, my BIOS doesn't allow me to make many alterations to the memory. In fact, all we changed was disabling the automatic setting for SDRAM configuration from 'by SPD' to user defined. Then, the values for the other entries (CAS latency, RAS to CAS delay, RAS precharge time, and SDRAM active to precharge delay) wouldn't give me the ability to make the changes the Falcon rep wanted to make. (He wanted to adjust the timing but I don't know his exact intentions... he would say something like 'It won't let you pick a higher/lower number?) I think the 4 settings were 2.5, T, T, and T. My new RAM is CL 2.5.

However, the good news is that I did not encounter another BSod -- the first time for that in several days. Not that the problem is licked, but maybe....

I moved the stick from the second slot to the third slot (the mobo doesn't care where the modules are placed) and the Falcon rep helped me return my system BIOS settings to the way they were configured originally. I had not realized that updating the BIOS had reset them to their default values.

I'm hoping that these 2 changes (using 3rd slot and BIOS tweaking) have made the difference. I wasn't able to reproduce the BSoD, but I also wasn't able to play WoW for very long...so maybe the system wasn't stressed enough. Running Memtest (test 1 only) with the memory in slots 1 and 3 did not produce the errors that had occurred before during test 1. But those errors had occurred only once anyway.

We may be getting a foot of snow today, so I'm working from home (can't be playing WoW)....and I'll test it out later this afternoon/evening.

I'll post again after that, just to let everyone know what happens (in case anyone is interested).

Again, thank you (and Matthias) for going out of your way to help me resolve this.

All the best.

 

[Gewesen]

Well, it looks like the stability issue might be fixed. I worked on the computer all day, then played WoW for about 2 1/2 hours (off and on). No BSoD!

Thanks to everyone for your help and patience!

 

[Fern]

Well good, glad to hear it! :thumbsup: