Is my Asus Z170 Sabertooth Mark 1 a dud?

Recently I bought new parts for my gaming rig, and carried over my existing drives & gpu. The new parts are an Asus Z170 Sabertooth Mark 1, an i7-6700K, and 2x8GB Corsair Vengeance LPX DDR4-2400. I initally I was getting WHEA_UNCORRECTABLE_ERROR messages in Win10, occasionally an IRQ_NOT_LESS_OR_EQUAL error, but I *thought* I had tweaked things and gotten the system stable. But now it's unable to do anything that seriously stresses the system without crashing & rebooting. I was able to do a little gaming, but if I try to run a video transcoding job with x265, it lasts maybe a couple of minutes before crashing. I tried the stress tests in both Asus RealBench and Intel XTU, none of them lasted more than a couple of minutes. Both the memory and CPU stress tests in XTU crashed after a minute. This was after I'd already tried downclocking the ram to 2133Hz, and capping the cpu at 4.0GHz.

I'm already regretting this choice of motherboard, I miss the software auto-tuning that my Z97-A has(the extensive BIOS options for tweaking cpu parameters are quite frankly confusing as hell), I *thought* I was getting the Sabertooth for added stability since I didn't need a significant overclock. And now that I don't have it, I'm at a loss as to whether I can just tweak the BIOS settings, or if I really do have some busted hardware.

Does anyone know a good way to try and isolate the problem?

Bios update, reset BIOS to defaults, don't touch anything, run stress tests. Does it pass?

Ocellaris said:

Bios update, reset BIOS to defaults, don't touch anything, run stress tests. Does it pass?

Click to expand...

First thing I did before installing the OS was to update to the latest BIOS on that board, it's currently running the latest version available(v1702). When I get home from work, I'll take it back to default settings, going as far as clearing the RTC jumper(even though the manual says you don't have to on a failed overclock). Once I do that, all I'll change is making sure the boot order is right. Thanks for the quick response. I hope I can get this thing running stable. I don't even need an overclock, the default cpu settings will be fine, and I'd like to get it to run the memory at its full 2400Hz, not just 2133.

I finally had a chance to try resetting the motherboard(CRAZY situation at work has pretty much eliminated any free time I have), after the system crashed on me again during the night(I usually leave music playing as I fall asleep). I went into the BIOS and chose "Load Optimized Defaults", then restarted. Still crashed on bootup. Tried turning off some of the auto-tuning and capped the cpu multipliers at 40, plus I disabled Turbo Boost. Same thing the machine won't boot into Windows. I get the startup screen, so it does POST, but it won't get into Windows itself. Most of the time it's the "WHEA_UNCORRECTABLE_ERROR" message, occasionally it's "IRQ_NOT_GREATER_OR_EQUAL" or whatever that one is. Once it crashed with "MACHINE_CHECK_EXCEPTION", which I hadn't seen before this morning.

Anyone have any ideas?

How does your RAM check out?

Run memtest?

mda said:

How does your RAM check out?

Run memtest?

Click to expand...

I'll need to find the time to do that, but yeah that's on the agenda. I'm on an emergency schedule at work right now, so I have VERY little time at home for the forseeable future.

I would take every component out of the board and clean and reinstall. Check the CPU for sure and see if you have a good seat on the cooler.

From there I would set the RAM timings by hand. I should run at 2400 spec without issue if the timings are correct.

Thanks for the tips, guys. I plan to keep troubleshooting this when time allows, but I'm stuck working 6x12's at work right now, as well as some time on Sundays :/

Did you do a clean install of Windows when you setup the new machine?

If not, I would try that as well.

You could have gotten a bum board or CPU.

What power supply do you have? It could be the power supply as well.

Make sure you don't have any shorts between the board and the case. Did you make sure that there weren't an extra standoffs when you changed the board?

cyclone3d said:

Did you do a clean install of Windows when you setup the new machine?

If not, I would try that as well.

You could have gotten a bum board or CPU.

What power supply do you have? It could be the power supply as well.

Make sure you don't have any shorts between the board and the case. Did you make sure that there weren't an extra standoffs when you changed the board?

Click to expand...

It's a brand new Corsair HX1200i PSU, and I did do a clean OS install. Part of this upgrade process for my machines was moving to Windows 10, so I did clean installs of Windows 7 on the new hardware configurations before requesting the free Win10 upgrade. The machine gave me trouble like this shortly after doing the installs, but after some BIOS fiddling that I can't remember the details of, I was able to get it running pretty stable(stable enough to run x265 encodes via ffmpeg overnight). But after several days these problems came back. I was eventually able to get it stable by lowering the CPU multiplier caps, manually setting some of the voltages(though AI Suite often showed voltages significantly higher than the manual voltage I'd put in), but I still couldn't stress the machine. Typical usage and even some gaming would be fine, but the encoding runs or the Asus RealBench stress tests would still crash the system. At some point last night, it crashed when just playing some audio files via MPC-HC, and now I can't get it to boot back into Windows.

Check using each memory module on its own in different slots. Keep default settings. IRQ BSODs can be memory related, while Whea can be memory or CPU side.

Raja@ASUS said:

Check using each memory module on its own in different slots. Keep default settings. IRQ BSODs can be memory related, while Whea can be memory or CPU side.

Click to expand...

Yep, might try each stick one at a time in the same slot as well and see if you find it that way.

OK, I finally had time to start running Memtest86, haven't had time to start moving the memory sticks around. With them both in the default slots, it appears to have gotten through 2 passes of all the tests with zero errors. I'm still going to try each of the memory sticks by itself in each of the slots, to see if it boots. Would running memtest on a single stick show errors even if I get none when that stick is running alongside its pair?

OK, this is....odd. If I run MemTest86's tests using all the cores, whether it's simultaneously, round-robin or sequentially, I get errors galore. If I run the tests using only a single core, it works fine for cores 0-3, but running it on cores 4-7(it's an i7-6700K), memtest86 reboots almost immediately after the test begins and starts running it on core 0. I've got to think it's something with the cpu at this point, yes? Or does MemTest86(v6.3.0free) always do that with hyperthreaded processors?

What did you find out? I'm about to purchase the Z170 Sabertooth board.

neokeelo said:

What did you find out? I'm about to purchase the Z170 Sabertooth board.

Click to expand...

Turned out to be a bum CPU. Got it replaced and the system is running like butter. I miss the software OCing in AISuite that the non-TUF models have, but I'm going to try the BIOS AI Tweaker this weekend to auto-overclock. For this machine though OCing isn't a priority.