Solved Random reboots, error reporting calls it BSOD, next gen hardware

[BotenRedWolf]

The nVidia driver procedure you sent me to.

 

[essenbe]

Darn. I thought you had already done that. But it's great to know that fixed it for you.

 

[BotenRedWolf]

I just find it really weird that changing a step or so in a driver re-install (which I'd done before), and blowing them out beforehand, can do. Granted, Windows still doesn't make a whole lot of sense to me, but hey, at least it works mostly like it should now. If it acts up again, I'll wake this thread back up.

Thanks for your help!

 

[essenbe]

You're welcome. The drivers are actually installed by Nvidia, but, all you are really doing is getting rid of all the older drivers, getting Nvidia to do a clean install and get rid of the remnants of older drivers and not installing anything that you don't need. One example is Nvidia installs an HDMI driver, which very few people use. Most motherboards have Realtek sound systems and include HDMI divers. Nvidia HDMI drivers can have a conflict with Realtek drivers. So, you are taking that possibility away. But, for whatever reason, it works. That's the way I install drivers every time and never have driver problems.

If you would, whenever you think your problem is solved, please mark this thread as solved. I'm glad everything is working for you.

 

[BotenRedWolf]

Marked as solved immediately after my last post. Strange that they would conflict like that though, or at least it appears that way to me...

 

[essenbe]

I noticed it was marked solved after I made that post. Guess I should pay more attention. The biggest thing about video card drivers, is removing as much as possible of the older driver before installing the new one. They even make programs to do that. 2 sets of almost any driver can cause conflicts, but it is most important with video cards, they can be very problematic as they are pretty complicated beasts. The HDMI driver, was just an example, and why Nvidia decided to add that is beyond me. But still it was only an example. The clean install thing is the trick. as much as possible, that tutorial kind of forces a clean install. Use it. Once you get used to it, it will be very simple and very quick. It works pretty well, especially for people who tend to have driver problems. That's why I use it and why I wrote it, to try to help people having driver problems.

 

[BotenRedWolf]

Alrighty, that was a short fix. It started doing it again, and also strange translucency and artifacting in game... grabbed some logs.

In the first one there were two distinct points while testing that where it dogged out bad.

The second one the FPS stayed high, but there was the artifacting and such present.

 

[essenbe]

Are you sure you set up the SLI correctly? And are you sure you have the cards in the correct slots recommended by Asus Manual? if not, please read this and see if everything is set up right Introduction to SLI Technology | GeForce

Failing that, I would break the SLI and try 1 card at a time by removing one and putting it through all the stress tests and making sure it is not the card. Do it in both slots And see what kind of performance and reaction you get out of it. Next, take that card out and try the other card in both slots Testing the crap out of it. This will test your motherboard slots as well as the cards.

We are reasonably sure that it Is not a driver problem and if you read the article and set up SLI correctly it is not the set up. The only thing I know is that if the SLI set up is right, the Drivers install is right, the only thing left is one card is defective, the motherboard slots are not right or the PSU is not putting out what is required to run SLI. As I see it, those are the only options left for us. If you like, I will ask someone more experienced with SLI to come here and take a look.

 

[ganjiry]

High BotenRedWolf,

I've just quickly skimmed over your thread, Am I right in thinking it's just Watch Dogs you're having the trouble with now? I've seen some of the same transparency/artifacting you mentioned while playing it myself but only the once. There are a few settings in the game that cause a lot of stutter and lag, I also found that running the Uplay launcher as admin stopped a lot of the game lock ups.
Are you running the game across all 3 screens? if so at what settings? Watch dogs is a killer of Vram and puts a strain on my cards..

 

[BotenRedWolf]

I did notice while leaving GPU-Z running on a different display while playing Watch Dogs that "PerfCap Reason" moves around in three different descriptions. Mostly: "thrm" which I take to mean thermal, as in running too hot. It's not a crippling reduction, in fact, it's barely noticeable.

The other two however, VRel and Util are harsher. VRel, which I think has something to do with the power supply lasts such a short amount of time that I can't notice a difference, while Util seems to show up very randomly and at times for a couple seconds at a time. I decided to force my GPU fans to 100% to cut out any thermal problems for the next tests... When Util is up for a while, that's when I get those horrendous frame rates and delays as I had originally described. There were times during "util" when the core and memory clocks would drop to almost their lowest possible states.

This may be out of the scope of this forum category, so either we split the thread or I ask for someone more into video hardware comes in.

 

[BotenRedWolf]

High BotenRedWolf,

I've just quickly skimmed over your thread, Am I right in thinking it's just Watch Dogs you're having the trouble with now? I've seen some of the same transparency/artifacting you mentioned while playing it myself but only the once. There are a few settings in the game that cause a lot of stutter and lag, I also found that running the Uplay launcher as admin stopped a lot of the game lock ups.
Are you running the game across all 3 screens? if so at what settings? Watch dogs is a killer of Vram and puts a strain on my cards..

It is only Watch Dogs at this point now, yes, but see my last post about GPU-Z readings and I have a feeling if I had any other heavy games like this they might perform funny too.

Definitely not running it across all three displays, since it's such a heavy game. When I have my display set for surround (a virtual single display) it disables SLI, which I need to run the game at a higher framerate.

Also... I'm not high... at least not right now.

 

[essenbe]

How does it do in Ungine Heaven at the extreme settings or Furmark ( http://www.sevenforums.com/tutorials/100356-video-card-stress-test-furmark.html )

 

[BotenRedWolf]

Heaven runs a good 40-50fps with only VRel. Furmark (mother of god, btw) VRel for about a half second, then Pwr, then after about 15 seconds, Thrm.

 

[essenbe]

Both tests are designed to get every bit of performance out of your cards they can. Furmark is absolutely brutal with video cards and, as I understand it, both AMD and Nvidia tried to outlaw it several years ago. They both put your card under more stress than any game ever will. If you can pass both of those tests, You have a game issue. It seems to me that your cards are doing much better than they were when this thread was started.

 

[BotenRedWolf]

No doubt they are, I haven't had it give me a driver recovery bubble or a random reboot since we did the fancy re-install. My only question now is what are those parameters on GPU-Z about Vrel and Pwr, and are they relevant to any problems I might stumble across elsewhere?

Mainly I'm asking does my PSU suck?

 

[essenbe]

Corsair is a top notch PSU and while it does happen, it is unusual for a Corsair PSU to be bad. I am running a Corsair and own 4 of them in other rigs. If you have reason to believe it is the PSU, fill out a ticket with Corsair. They have a good warranty. That being said, Watch Dogs has caused quite a few people problems, some with better hardware than either you or I have.

 

[BotenRedWolf]

I meant is 750W enough for all my hardware? I didnt plan SLI when I picked the PSU.

 

[essenbe]

From what I find the estimated max power draw for 2 X 760 is 343 Watts. The recommended power supply is 700-800 Watts. Source

 

[BotenRedWolf]

What voltage rail do the cards source off of? Another thing I was thinking when it had Vrel errors is if Vrel is a voltage error, then maybe too much current through the source wires (causing a voltage drop due to low gauge wire) heading into the card from the PSU. I have additional leads since it's a modular supply.

 

[essenbe]

Try another lead if you like, it can't hurt. But, GPUz, like most other software has to fix a reason, and is not always right. It has to blame something. Consensus of opinion is when your cards are maxed out, they cannot continue boosting to infinity, basically what it's telling you is your card has boosted to it's max capacity. That is my understanding on the readings I have done. GPUz will tell you on the main page what your max boost is. If you are very close or over that, you have reached the max boost without overclocking it.