Solved Random Freezes, BSOD, error 0x116

[Baasha]

RAID: Marvell* SATA Driver should help.

Okay, some updates.

That RAID Controller driver is mv91xx but my RAID Controller is mv61xx. The ones in that link do NOT work. I think I have the latest ones for the mv61xx controller; I got it from station-drivers.net.

I set everything to default, no OC at all, and the system still crashed when I played that game (Trine 2). I tried another game, same thing. I tried reinstalling NVIDIA driver, same thing; I even rolled it back to 296.10 WHQL to no avail.

What is so annoying is that when I do 3-Way SLI, with BOTH the CPU & GPU OC'd (or even default), there are no crashes! I was able to play Trine 2, Metro 2033, and other games without issue! This time, I even had the three adjacent PCI-E lanes occupied; lanes 1, 3, and 5!

I went ahead and got an RMA with ASUS. I just haven't shipped out the board to them yet. What do you think I should do? I think I've exhausted all troubleshooting methods. Should I just go ahead and RMA the board?

 

[writhziden]

Sorry the link was for 91xx; it would make everyone's lives so much easier if providers would list the .sys file... Marvell hard disk controller drivers seem to be difficult to find, too.

I do not know if that site you found the drivers from is trustworthy. I will post a thread in the drivers section of the forums and see if our driver experts can help me find a trustworthy, working 61xx driver.

As for the motherboard, it is your call. Going back through the thread, I see you did a repair install at one point. Have you done a clean install yet to see if that would resolve it. You mentioned re-installing Windows, but I cannot tell if that meant you clean installed it or you were referring to the repair install.

If you http://www.sevenforums.com/tutorials/219487-clean-reinstall-factory-oem-windows-7-a.html#post1839164 and then only install the applications causing crashes, that would help determine whether your Windows/software is magnafoozled or it is hardware.


EDIT: I started a thread in the drivers forum: http://www.sevenforums.com/drivers/...-sys-marvell-raid-controller.html#post1944902. I found out the station-drivers.com site does provide reliable drivers. Further research I did on my own suggests the latest driver on that site may be a beta, though. ASUSTeK Computer Inc. -Support- Drivers and Download P5Q Deluxe -> SATA -> Beta Version 1.2.0.8300

If you used 1.2.0.8300, you may consider switching to Version 1.2.0.8000 to see if it provides more stability.

 

[Baasha]

Sorry the link was for 91xx; it would make everyone's lives so much easier if providers would list the .sys file... Marvell hard disk controller drivers seem to be difficult to find, too.

I do not know if that site you found the drivers from is trustworthy. I will post a thread in the drivers section of the forums and see if our driver experts can help me find a trustworthy, working 61xx driver.

As for the motherboard, it is your call. Going back through the thread, I see you did a repair install at one point. Have you done a clean install yet to see if that would resolve it. You mentioned re-installing Windows, but I cannot tell if that meant you clean installed it or you were referring to the repair install.

If you http://www.sevenforums.com/tutorials/219487-clean-reinstall-factory-oem-windows-7-a.html#post1839164 and then only install the applications causing crashes, that would help determine whether your Windows/software is magnafoozled or it is hardware.


EDIT: I started a thread in the drivers forum: http://www.sevenforums.com/drivers/...-sys-marvell-raid-controller.html#post1944902. I found out the station-drivers.com site does provide reliable drivers. Further research I did on my own suggests the latest driver on that site may be a beta, though. ASUSTeK Computer Inc. -Support- Drivers and Download P5Q Deluxe -> SATA -> Beta Version 1.2.0.8300

If you used 1.2.0.8300, you may consider switching to Version 1.2.0.8000 to see if it provides more stability.

Thanks for the updated links. I think the RAID controller is working fine; I've never had any trouble with it.

I am going to RMA the motherboard. Yesterday, the computer crashed even with 3-way SLI after about 15 minutes of game play. So, instead of crashing right away (with 4-Way SLI), ti took a little longer to do so.

I did do a clean install of Windows 7. In fact, I did it twice. I reset the CMOS and reflashed the BIOS and then installed Windows 7 after secure erasing the SSDs. None of it helped.

Hopefully the new board they send me will work well.

Thanks for the advice and help.

 

[writhziden]

You're welcome. Sorry we couldn't find a simpler solution. Let us know when you get a replacement motherboard and whether it resolves the issue. Best of luck it does!

 

[Baasha]

You're welcome. Sorry we couldn't find a simpler solution. Let us know when you get a replacement motherboard and whether it resolves the issue. Best of luck it does!

Okay, I got the motherboard back today from ASUS. The thing is, they sent me the SAME board again!

When I called their customer service and spoke to them, they said that all they did was reset the CMOS battery and "reset everything to default"!?!

The manager said he didn't have more information on what they actually did!!

And, sure enough, after spending about 2 hours setting everything up, the problem is STILL THERE! BSOD with 0x116. Heaven 3.0 doesn't even run for more than 10 seconds before locking up and BSOD'ing (again, 0x116). I am really starting to wonder if it is one (or more) of the GPUs.?!? Even though I've tested all of them individually, the BCC 0x116 is most likely the GPU(s).

I am going to be calling ASUS but what else do you think I can do at this point??

 

[writhziden]

If you run all three boards in the system together without the SLI bridges, do you still have problems? Have you tried different SLI bridges?

 

[Baasha]

If you run all three boards in the system together without the SLI bridges, do you still have problems? Have you tried different SLI bridges?

I haven't tried running all 4 or 3 of the GPUs without the SLI bridge. I'll try that but I don't see how that would help.

I did, however, use 2 different SLI bridges; one is a 3-Way SLI bridge and another a 4-Way SLI bridge. Last night, I had multiple crashes with both 3-Way and 4-Way SLI. So, I'm pretty sure it's not the SLI bridges.

I am now testing the system with one GPU (again). I reinstalled the drivers (301.42 WHQL) and ran Heaven 3.0 for multiple runs and it ran fine. I haven't tried any games on it yet but will do that later today.

Although, I am going to call ASUS on Monday to get a new board. I can't believe they sent me the same board after resetting the CMOS and the BIOS - that doesn't address any hardware issue with the board. I wrote a clear note that they should test 3-Way and 4-Way SLI and I doubt they did that. There was no note with the board when I received it stating what they actually tested! Very poor service IMO.

 

[writhziden]

Yeah, I've heard some bad reports with a number of motherboard companies. It seems to depend on the actual people involved in the service. Sometimes you get technicians that care, and other times you get people who just want to get it done and shipped out so they can move on to the next problem. It is unfortunately a drop in the bucket...

Let us know what ASUS says and whether you determine anything by running the cards together without SLI enabled. The point of that test is to see if you still crash with the cards in even if they are not running collectively.

 

[Baasha]

Yeah, I've heard some bad reports with a number of motherboard companies. It seems to depend on the actual people involved in the service. Sometimes you get technicians that care, and other times you get people who just want to get it done and shipped out so they can move on to the next problem. It is unfortunately a drop in the bucket...

Let us know what ASUS says and whether you determine anything by running the cards together without SLI enabled. The point of that test is to see if you still crash with the cards in even if they are not running collectively.

Okay, so I tried what you said; running all the cards without the SLI bridge. The system seems to be running fine in that setup. When I tried SLI even with two cards, Heaven 3.0 locked up within a few seconds and the computer froze forcing me to hard-reset. Without SLI, I was able to run Heaven 3.0 for an hour maxed out (of course, only 1 card was being stressed) and also play some games without issue.

So could it be some issue with SLI? I've used 3 different SLI bridges and all of them cause problems so it's not the bridges themselves. When I use 3-Way or 4-Way, I'm having to kind of squeeze the cards together to get the bridge(s) to install properly. They don't go into the slots correctly otherwise. Could this be causing the issues? I doubt it though because when I used the SLI connector (for 2-way), which is an orange flexible connector, there were no issues of installation but the system still crashed. It certainly seems like SLI is totally screwed up but how do I make sure? Even if it is the SLI at fault, how do I fix it?

 

[writhziden]

Have you tried the 2-way SLI connector on all configurations. There are three possible setups with it. If we label cards 1-3, you can do 1 and 2, 1 and 3, or 2 and 3. See if all yield the same results of freezing. We need to determine if maybe one of the cards has a bad SLI contact.

 

[Baasha]

Have you tried the 2-way SLI connector on all configurations. There are three possible setups with it. If we label cards 1-3, you can do 1 and 2, 1 and 3, or 2 and 3. See if all yield the same results of freezing. We need to determine if maybe one of the cards has a bad SLI contact.

Okay, another update. I tried exactly as you suggested and I think we are getting somewhere (although I am not holding my breath).

I first tried GPU # 1 and GPU # 3 in SLI with the 2-Way connector. NO ISSUES AT ALL! I ran Heaven 3.0 through about 5 loops and benchmarked it as well; no hiccups at all. Then, I played Trine 2 MAXED out with the GPUs (#1 & # 3) being used 99% while the other two were @ 0%. Again, NO ISSUES AT ALL! Then I tried CoD: Black Ops maxed out, again, no issues!

Then, I tried connecting GPU # 1 and GPU # 2 with the 2-Way connector. About 30 seconds into Heaven 3.0, the system crashed! I then shutdown the computer and tried the other set of connectors (each GPU has 2 SLI connectors - I tried the ones closest to the bracket first and then the ones towards the center of the card) to no avail; Heaven 3.0 crashed within a few seconds as well!

I then tried connecting GPU # 2 and GPU # 3, but had my DVI-D cable connected to GPU # 1. This would not let me turn on SLI (since #2 & #3 were the GPUs with the SLI connector). I then shut down the computer and plugged in the DVI-D cable to GPU #2. No video signal at all! I heard the computer boot into windows but the monitor was in sleep mode (flashing orange). I guess the video signal was being put out only through GPU # 1. I tried both DVI-D ports on GPU # 2 to no avail; no video signal at all.

So, that's what I've done so far. It seems like I HAVE to use the DVI-D cable on GPU # 1 to get the video signal so I am not sure how I can test, for instance, GPU # 2 and GPU # 3 or GPU # 2 and GPU # 4 (or even # 3 and # 4).

What should I do next to narrow it down further?

Oh, btw, I re-tried 4-Way SLI for fun to see if luck was on my side but as all things in life, nothing is that easy!

I've attached the dump files so please let me know what it says (I would wager the NVIDIA driver and/or GPU since the BCC was 0x116).

 

[writhziden]

Yep, sounds like card #2 may be an issue. Are you not able to remove card #1 and have only #2 and #3 in the system?

 

[James7679]

If I may jump in here? Have you marked the "#2 GPU" that you suspect as bad? I see you are trying to run, up to, 4 way SLI? Remove the #2 from the loop, and move the other two over, and retry.

 

[Baasha]

Yep, sounds like card #2 may be an issue. Are you not able to remove card #1 and have only #2 and #3 in the system?

I can remove the cards from the system. Although, now I've run into some really weird issue.

This morning, NVIDIA released their 304.48 Beta drivers. I uninstalled the previous drivers (301.42 WHQL) and installed these. I was in 4-Way SLI and after setting it up, I ran Heaven 3.0 and as usual, it crashed about half-way into the test.

Then, when the computer rebooted into Windows, I disabled SLI and wanted to shut it down to remove the cards and try what you suggested. The computer DOES NOT SHUT DOWN!

It keeps restarting instead of shutting down! This is so weird I've NEVER had this happen before! I clicked "SHUT DOWN" and even created a 1-second shutdown shortcut on the desktop to no avail; it keeps rebooting every single time!

HELP!

 

[Baasha]

If I may jump in here? Have you marked the "#2 GPU" that you suspect as bad? I see you are trying to run, up to, 4 way SLI? Remove the #2 from the loop, and move the other two over, and retry.

Yes, I have marked the #2 GPU. As mentioned above, I was going to remove it and try what was suggested but I can't get the computer to shut down! Of course, I could do a hard-reset but this is ridiculous! Now, I can't even shut down the computer!

 

[James7679]

If I may jump in here? Have you marked the "#2 GPU" that you suspect as bad? I see you are trying to run, up to, 4 way SLI? Remove the #2 from the loop, and move the other two over, and retry.

Yes, I have marked the #2 GPU. As mentioned above, I was going to remove it and try what was suggested but I can't get the computer to shut down! Of course, I could do a hard-reset but this is ridiculous! Now, I can't even shut down the computer!

How are you attempting to shut it down?

 

[Baasha]

If I may jump in here? Have you marked the "#2 GPU" that you suspect as bad? I see you are trying to run, up to, 4 way SLI? Remove the #2 from the loop, and move the other two over, and retry.

Yes, I have marked the #2 GPU. As mentioned above, I was going to remove it and try what was suggested but I can't get the computer to shut down! Of course, I could do a hard-reset but this is ridiculous! Now, I can't even shut down the computer!

How are you attempting to shut it down?

I clicked on the Start button and then clicked on the "Shut Down" button. Anyway, I restarted the computer using Safe Mode and then it was able to shut down properly.

I then removed 2 of the GPUs in order to test GPU # 2 and now the computer doesn't even POST!

I moved GPU # 2 to Slot # 5 (PCI-E) and the computer still doesn't POST. So I removed GPU # 1 but kept GPU # 2, plugged in the DVI-D cable into that card and the computer started fine. I ran Heaven 3.0 for 10 loops and it passed without any issue even OC'd. The moment I plug in another card, either Slot # 1 or # 3, the computer doesn't POST. Seriously, what the flying *%$& is going on? By the way, did you have a chance to look at those BSOD dumps? What do they say?

What should I do now that the computer won't even POST with two GPUs installed?

 

[James7679]

Okay, let's try something. Please follow the instructions on this page. When finished, leave all of your GPUs out of the PC, except for the one in the first lane. Please double check all of your connections and be sure that everything, including the RAM, is seated as it should be. After doing this, reboot the system , go directly into the BIOS and select, "LOAD FAIL SAFE SETTINGS". I don't care about any of your overclocks. Leave the machine running as factory as possible for now. Select F10 and select "Y" hit enter. IF the machine boots into windows fine, mark the card as good. If it does not boot into windows, just set it aside. Power down, repeat the WHOLE process until all of your GPUs have been tested in the FIRST lane. If you find that you are unable to mark any of the cards as good, RMA the MoBo. Post back with your results.

 

[Baasha]

Okay, let's try something. Please follow the instructions on this page. When finished, leave all of your GPUs out of the PC, except for the one in the first lane. Please double check all of your connections and be sure that everything, including the RAM, is seated as it should be. After doing this, reboot the system , go directly into the BIOS and select, "LOAD FAIL SAFE SETTINGS". I don't care about any of your overclocks. Leave the machine running as factory as possible for now. Select F10 and select "Y" hit enter. IF the machine boots into windows fine, mark the card as good. If it does not boot into windows, just set it aside. Power down, repeat the WHOLE process until all of your GPUs have been tested in the FIRST lane. If you find that you are unable to mark any of the cards as good, RMA the MoBo. Post back with your results.

Okay, I did exactly as you suggested. Tried all four GPUs in Lane # 1. All of them booted into Windows fine without issue. Of course, that reset my CPU OC as well since the BIOS was set to defaults. After this test, I put the (originally thought-problematic) GPU # 2 in Lane # 1 and GPU # 1 in Lane # 5 and used the 2-Way SLI bridge. At 910MHz, the system crashed. GPU # 2, btw, by itself, passed at that frequency without issue! So, only in SLI, does GPU # 2 in conjunction with any other GPU crash in OC mode. Heaven 3.0 passed without issue at stock clocks (855MHz) with GPU # 2 and GPU # 1 in SLI. I tried OC BIOS mode (specifically made for the GTX-580 Classified) and the system still crashed. I think GPU # 2 is the weak link in the chain.

Now that the system works in SLI, should I try 3-Way SLI with the other three supposedly "good" cards? That way, I can try to narrow it down even further.

*EDIT: I tried 3-Way SLI and it's hopeless. It crashes as soon as I get to the desktop. This is with GPU # 1, 3, and 4 btw.

 

[James7679]

Baasha, were you able to get this resolved?