Help trying to get RAM stable on Asus 650i

[Erydin]

I am unable to get this board stable. I posted in another thread trying to get some help to flash the bios from 0302 to 0501 (supposedly a better bios). But I am not convinced that is going to fix my problems. System as right now is:
- Board: P5n32-E SLI Plus (650i)
- Processor: E6600
- Ram: Corsair 6400C4 (4-4-4-12T, 2.1V)
- Graphics: 8800GTX
- PSU: Corsair/Seasonic 620W
- O/S: Vista 32bit

I am getting lots of BSOD. The BSOD error always seems to be a 0x000007f, followed by some other numbers in paranthesis the first of which is usually something close to 0xc0000008. The other error messages listed after this cruise by to fast to get all of it but they are also different depending on the crash. It would seem from what I can find that these error messages generally relate to memory but can also be hardware (Microsoft was not a lot of help on what they mean).

Now, I have 4 1gig sticks of this memory. If I place all 4 in at the same time it will POST but windows will not load. 3 sticks the same. 2 sticks in (regardless of which slots they are in) will boot runs seamlessly, UNTIL, I try to run a game that uses lots of system resources. I can get a BSOD of death immediately after signing into BF2142 and in under 60 seconds from Vanguard, every time. WoW strangely enough runs without a hitch. I can run with one single stick in and these games bog down bad but they do not crash with only one stick in.

What I have done for testing next,
- Memtest86 (v.165 and 170): Two memory sticks, no problems detected. I have taken these two sticks put them in the first two slots (A1 B1) and moved them around in what believe is called dual channel, the first and third slots and the second and fourth (A1-A2 and B1-B2). Regardless with two memory sticks zero errors from Memtest. 3 sticks, some errors starting in Test 2, there is 10-20 seconds between each error, I am guessing when it tests the third stick it fails and the other two are passing. 4 sticks in, total 100% failure of all tests in Memtest.
- Orthos: It is in windows so I have to be able to boot, two sticks with blend not a problem detected ever. 1 stick the same. Placement doesn't seem to matter, they all pass, never error. However, I noticed one strange thing here, when running the ram test only, I look over at my Vista gadgets and the one that rates CPU and Mem Usage shows the CPU at 100% but the Mem doesn't go up at all. Blend test shoots both the Mem and CPU to 100% though.
- 3dMark06: Just ran this to make sure graphics card was working fine. Decent scores, never has problems, finishes the benchmarks doesn't have any errors that I can see.

Now I ran DxDiag just to look at things and something very very odd pops up here. Under display it correctly lists my 8800GTX. However, if I have two sticks of memory in it shows the the Memory under the graphics card to be 1499MB. Basically double what it should be. If I run just one single ram stick it shows just under 768MB which is what the card should have. For some reason when I put in two sticks of ram the dxdiag is thinking my graphics card has double memory... is my system then thinking I am running 8800GTX in SLI perhaps?

I have tried so many settings in bios my brain hurts. I have forced the timing to 4-4-4-12-2T, 4-4-4-15-2T, 4-4-4-Auto-Auto, all options Auto, I have overvolted to 2.1 and left it at Auto, tried 2.0 and several other voltage setting as well. I have set the SLI memory option to Expert, Disabled, and CPU0%.

I am really at a loss now why my system is still getting BSOD. I have also OCed the processor to around 3.1Ghz and it runs just fine, passes all the tests just like it does at default and still BSOD at the same points. I have been tooling around with this board for well over a week, several hours a day after work and just can't seem to figure out what is going on with it. I have checked Asus' website as well as Corsair's and there are plenty of people that are running my exact ram with this board with no problems (the problems they are reporting is when they try for 4gig... hell I would be happy to get it stable at 2gigs right now). Anyone's suggestions would be greatly appreciated,

Erydin

[Yellowbeard]

I would suggest 2 things. First, make sure you have REALLY good cooling on your NB as that thing gets HOT HOT HOT. Mine is water cooled now but I also used a Swifty on it. Second, you will need to raise your NB voltage 1 or 2 bumps with 4 DIMMs.

[Erydin]

My system is also watercooled. Two DDC2 pumps with Petras tops, two PA120.3 rads, in my loop I have the graphics card, CPU, and NB. Core Temp shows under full load (Orthos blend running for a good long time), even when OCed to 3.1Ghz, max temp I see is 44 degrees, it idles around 32-33. I have tried NB voltages of everything from 1.30 up to 1.55. It seems to default, from what I can see, at about 1.30.

I also was able to get WoW to crash to BSOD last night. I opened several internet windows, left these on different pages, booted up WoW, parked my character in this game in a busier area with other players coming going and lots of other things going on at this place. I then minimized WoW. Vista gadgets showed my RAM usage sitting at about 48%. Opened another internet window and grabbed a site with lots of animations and the like on it. Page nearly finished loading and BSOD happened. Makes me think when my computer uses over 50% of the Ram possibly I am getting BSOD... but its hard to say if this is for sure what is happening. Any other suggestions or things to try?

Erydin

Edited for spelling

[Yellowbeard]

Sounds like you have covered most of the bases. I am running 4 DIMMs on that same board now with a 7950GT/OC SLI setup and the same PSU. Your RAM tests out OK so the only major difference I see is Vista as I am using XP SP2. Is there any way at all you could test this setup with XP? Vista is still a mystery to me but that's my only idea for now.

[Erydin]

Lets pretend for a second that I had a hypothetical friend who had all the same problemds I did. Now lets say this friend was looking at his set up again and realized he had placed his 8800GTX card in the lowest (i.e. farthest from the processor) pciE x16 slot. Now why he did this I am sure is because he was more concerned of how he was setting up his water loop and the lowest slot offered the best loop configuration (i.e. a gentle loop from the NB to the graphics card with the smallest amount of stress from the tubing pulling on the NB block). Now this friend was thinking of the loop configuration and not that you should put a single graphics card in the highest slot on the board not the lowest. Now this friend being all "Dee Dee Dee" realized his mistake. Could this be causing all the problems listed above... for my hypothetical friend of course.

My only problem with saying this would be the problem is that the system passes 3dMark06 with no problems. Of course the DxDiag reporting twice the memory that should be on the card leans me to thinking that card placement could be indeed what is wrong. I would like to avoid moving the card... err I mean having my friend move his card that is, unless I am sure it is causing the problems since the loop has to be drained, tubing removed/recut/remounted, etc... basically its a big pain in the butt.

So, has anyone else placed their card in the lowest slot on an SLI and if so have you exprerienced problems such as these or did it work fine? Thanks in advance.

Erydin

[Erydin]

Yellowbeard, XP may be a problem to get hold of for me at the moment. I can try to get a copy but it could take a bit of time yet.

Erydin

[Yellowbeard]

I have no idea what effect the card placement may have on this. But, it seems like both the OS or the card placement will be a pain to test.

Here's 1 thing you could do if you still have your stock fan for your card. Remove the water block from the video card and leave your water loop intact. Put the stock cooler on the video card and test the system with the card in the top slot. That way, you can test it without opening your loop.

[Erydin]

Problem Fixed. Going to post on what I did to get it working again just in case someone else has this problem at some point also.

To start with it was not a graphics card or RAM problem, so a moderator may want to rename the post and move the whole thing to Motherboard section.

Problem: Chipset Drivers
Solution: Downladed the updated ones (duh)

It appears my chipset drivers were not getting along with something else out of my new build. Guessing it was VISTA since part of the updated chipset drivers were some new beta VISTA drivers.

Since I have done this, I can't get a BSOD no matter how hard I try, and I tried really hard to test the stability. Its a good day.

As a side a note this was after I tore my entire loop down and moved my graphics card, no fix doing this. Got hold of an XP copy, whiped my drive and reinstalled this OS, no fix here either. Was looking at my VISTA copy sitting next to my computer and it just occured to me that maybe I could try updating chipset drivers since I had not much left to do. And it worked great. Thanks for the suggestions and help :)

Erydin

[Yellowbeard]

Glad to hear it, at least the solution was free.