Asus Maximus Hero VIII /w Asus Geforce GTX 1080's in SLI

[NostalgiaBytes]

Hey All,

First time poster here, hope you're all well, and I'm hoping you'll be able to help me with an unusual issue I've been having of late too. So my system specs should be in my profile, If you cant see them ill update this thread as necessary. Anyway onto the problem -

I built my system about three weeks ago, hadnt had any issues baring the fairly common memory not working on XMP etc, thankfully resolved with the release of Bios 1902. It's been working perfectly effectively. That is, until I installed my secound Asus Founders Edition Geforce GTX 1080. I now find my self in a situation where for some inexplicable reason my CPU will throttle to 800Mhz and stay there. This can happen 20 minutes after boot, or work perfectly for 5 hours (tested with stop watch, OCD me). But when it does throttle down to 800Mhz, it wont budge from 100 x 8.0 (800Mhz) no matter what I try to do within windows once its happend. It can occur if the machie is left idling all day, or even in games and applications, it has no ryhme or reason it seems.

I have to perform a shut down and reboot, and low and behold, its fine again. She'll trundle along happily, working, until the random afflication occurs again, which is anytime, sometimes not at all.

If I yank out the second 1080 GTX, well, all is well again, tested with the machine on 3 days straight. If I swap the first card out and run on just the second card (incase a fault with that card) again, no problems.

As for the various other things I've tried - I've tested with C-States and Intel Speed Step disabled. No Dice. Problem still rears its ugly head.

So I'm at abit of loss as to what the problem might be, after all, if i pull the second GTX the PC is fine, so what would I RMA the CPU? The motherboard? Is it a case of the Hero VIII cant handle every PCI slot being filled, particularly with two 1080's and an AC1900 wifi card, soundcard etc. Or cant the CPU handle two 1080s? I feel like this all should work together, but if i need to upgrade to the Alpha or Extreme I will, just want advice, help before I head down that route.

Thanks for any help and guidance you can give me on this, it's very much appreciated.

[MeanMachine]

You should conduct a series of stress tests for instability using AIDA64 to determine Voltages and temperatures whilst under load.
Also download HWMonitor and put them on your desktop side by side.
This way you can look for Vdroop and max Temps in HWMonitor whilst the test are being run on CPU and FPU.
You can also individually test your GPU cards with AIDA64 and Your DIMMs.

Have you OC your CPU or made any changes in Bios from default. ?

I'm a little concerned regarding the H55 being able to cool efficiently.

[NostalgiaBytes]

Hey, I've got HWMonitor and CPUZ set to load at start up, idle temps of the CPU are 19 - 25c depending on ambient temps (live in Uk) load temps whilst running real bench, max out between 65c and 70c

I'm nigh on 99.9% sure its not a temprature issue due to the fact, if I yank the second GPU, running the rig on just one of the 1080's (doesnt matter which) the system works happily, all day every day.

Also I dont believe the extra GPU appears to put any extra load on the CPU specifically, if it does, voltages, temps dont appear to show any difference. Now as for the PCI Lanes which are full, well thats a different story.

It's a weird one that's for sure, and at the moment I'm running my shiny new machine minus one of its £600 components, which is abit frustrating.

I'm gonna run some more tests over the weekend, but I think its heading towards me replacing it with a more competant motherboard. Thinking either the Alpha or Extreme. Utilise the Hero in a smaller less demanding setup.

[Chino]

Try running the High Performance profile for your Power Plan.

[NemesisChild]

Have you tried reverting the bios to default settings and testing it from there? Or cleared the CMOS lately?

Running SLI in a mid-size case with a H55 cooler....I think you have a CPU heat throttling issue.

[NostalgiaBytes]

So following a weekend of rigorous testing, think the issue lies with the motherboard, can’t explain it otherwise.

I've ruled out temperature, for two main reasons, running the rig with the side panels off, in my offices comms room has the entire rig running at a frigid 17c and 28c maximum under load. Also if the rig is left in a powered down (soft off) state overnight and you turn it on next day (so powered down for several hours), it will load with the CPU throttle locked to 800Mhz. If I was to flick the rocker or disable c-state, ErP modes it boots just fine. However as stated before, it'll throttle lock to 800mhz anywhere from a few minutes after boot to several hours. Other issues I've noted is just a general shaky level of stability, USB devices disconnect and reconnect randomly when two cards are installed, exiting 3D applications, sometimes cause the system to freak out and grind the desktop to a crawl.

The only common theme, is that if I yank that second GPU, the entire system operates normally, in fact id even go so far as to say it’s faster, smoother.

In summary though, with a second GPU installed, it doesn’t appear to matter how or what I try, it eventually flakes out. Be that the system with just two GPU's and nothing in any other PC Slot, running one 16Gb dimm as opposed to two, or running with optimise defaults or customised settings in the BIOS.

The fact from a soft off boot it will without fail, every time, lock to 800mhz even if the CPU temp never went above 20c from cold boot - but I do the same with one GPU its fine, after all my testing the only thing I can think, is there something iffy with the board, but it only appears when it's asked to run two GPU. So I'm thinking maybe the best course of action is to replace it with a different series, was thinking trying the Sabretooth Mark 1 unless anyone has reason to avoid that board. I dont see a reason to RMA the VIII Hero board when I can utilise it elsewhere as long I ensure I dont use it for SLI. Just hope this finally resolves the issue. Because after replacing the PSU (after thinking it was a power issue) trying everthing I can think of, I'm getting abit tired of it now.

[Subsider]

did you try running the High Performance profile for your Power Plan as Chino suggested ? are you using a UPS ?