Machine suddenly shut itself off during Daz Studio render
nomad-ads_8ecd56922e
Posts: 1,873
Okay, this is flipping strange. I set up a render with one of KindredArts' clothing decals, and a short ways into the render, my machine just suddenly SHUT OFF without warning. I have a fairly beefy power supply, a Corsair RM750, and haven't had any trouble like this since I got it maybe 4 or 5 years ago. (It can't be older than that, since the box has a 2014 copyright date on it.) This wasn't an overly complex scene, either, basically 3 people and a bus in front of a multi-plane-diorama background. Any idea why my machine would suddenly do that over something like this? 0o And where would I look to see if it what caused it in DS was logged?
Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Well, when I see your description my first thougth is overheating. Could be either the CPU or the GPU, but most likely the CPU.
Have you any ways to monitor temperatures?
It could just be that there has been built up some dust inside your computer getting it harder to maintain acceptable working temperatures.
In DAZ studio Help tab see logfile
Also check event viewer under Windows start control panel administrative tools advanced
I have the FireStorm desktop app that came the makers of my card (a GeForce GTX1060), and also a more general hardware monitoring app called HWiNFO64 that also puts up a window that monitors the operating temperature of all the componants (ram, cpu, video card, HDDs, etc) but I wasn't running either of them today. Last time I looked at them, earlier this week, the temperatures of everything were within normal normal ranges.
MInd you, I've also been having issues with Win10 1809 giving sudden hangs out of the blue, but every time I had those apps running and one of those hangs happened, all the componants were still showing as within their normal temperature ranges. This hang problem showed up out of the blue right after Windoze update to 1809. and everything was completely problem free before that update. I've been getting help on one of the big Windows tech-help forums, but none of the actions they had me take has fixed it. The only remaining one left is a total reinstall, which I Am NOT going to do. 0o
unforced shutdowns can have a number of causes.
Heat is the usual culprit. When was the lasttime you cleaned the dust out?
A single unforced shutdown could also just be that Windows crashed in a particularly spectacular way. If it doesn't happen again I'd leave well enough alone.
A bad memory stick can cause this as well. If you're not overheating you need to run a test on your RAM. memtest is the usual utility for this. Follow the instructions on their site.
A power spike could cause this as well. Whether your PSU is "beefy" isn't the issue. It could be failing or the filtering hardware on your surge protector could be failing.
After that you run into the bad stuff like a motherboard or CPU failing which could be an absolute PITA to pin down.
Just to be clear, you are saying that the PC completely shuts down, right? Not DAZ Studio or Windows, but the whole machine as if you had turned it off? If that is the case, I would be exhausting all hardware problems first, as a sudden shutdown suggests that. Have you been messing around inside the machine recently - hardware changes, etc.? If the machine will boot and run, then randomly shut down, I would suspect a short. If the shutdown only happens under stress (high load), then cooling might be the issue. Problems with the PSU could also cause such behaviour and it would be worth watching the voltage regulation on the different rails while the machine is under load to ensure they remain within parameters. PSUs, even good ones, will not necessarily last forever. If its not exessive heat, a failing PSU or a short, the only thing left that would cause that sort of behaviour that I can think of would be a failing component on the motherboard, which is difficult to diagnose, unfortunately.
I just had a few additional thoughts - troubleshooting is always doubly difficult by remote control.
Is your PC plugged directly into a wall socket, or is there an intervening power bar or UPS? If the latter, then temporarily, you could try plugging the machine directly into the outlet, or use an alternative power bar/UPS if available and see if the behaviour persists. Finally, you could also try plugging the machine into a different outlet, preferably on an entirely different circuit, and see if that solves the problem.
An intermittant short is very unlikely. It is possible, solder connections do sometimes crack with age and surface mount components can go bad, but in better than 2 decades working with servers I've seen, IIRC, 3 of them.
Generally if there is a short it doesn't just crash the system but burns up one or more components, I've seen actual fires from shorts.
Yeah, the machine just simply shut itself off as if I'd yanked the plug out of the wall. FOOOoooooom!!!!
My machine is on an APC Back-UPS XS 1500M, and I have the PowerChute monitoring app installed on my machine, but hadn't really looked at the app in ages. That pretty much rules out voltage spikes or the like. The UPS is rated for up to 900 watts, and my machine is using a 750 watt PSU. That said, I did add a second HDD and switched to a more powerful video card since I got that PSU, so that might have pushed the peak wattage up. On the other hand, PowerChute is showing me at using 126 watts of power at the moment, tho I assume that also includes the two LCD monitors and whatnot. Mind you, that is with me simply running my usual apps and desktop programs (web browser with a gazillion tabs pinned, several seperate chat programs, Thunderbird email, etc.) and without DS running at the moment. I suppose I should park my PowerChute window on the 2nd monitor along with HWiNFO64 and maybe FireStorm (tho HWiNFO64 also monitors video card heat and fan performance) and watch those like a hawk next time I try this render.
The UPS isn't all that old, its from maybe late last year or so. (Goes and checks.) Actually, I recieved it November 30, 2018.
As for the PSU, I got that in November of 2015.
Well if event Viewer and your monitoring software isn’t telling you anything my guess would be it’s power related as anything else shutting it down would leave some indication, it could even be a dodgy cord.
Unfortunately, Wendy, monitoring software will only tell you so much about certain key components. There are hundreds of components mounted on a motherboard, for example, and monitoring software will tell you very little about the status of those individual components, so we are left with trying to identify issues from clues. Once the usual suspects are eliminated, troubleshooting becomes much more difficult.
I would agree with your points about shorts, kenshaw, with perhaps one exception: mechanical switches. Since the latter are operated intermittently by the user themselves and also subject to wear and tear, they can develop problems and shorts that certainly can be intermittent. Then there is that pesky ghost in the machine.
It does?
When was the last time you made sure your UPS was working as intended? Is your monitoring app working as intended?
Folks make troubleshooting harder when they rule stuff out because... Rule stuff out only after checking.
Thing is, I haven't had ANY noticible trouble until a couple weeks ago when Win10 updated itself to 1809.... and then I've had random hang after random hang, where everything just freezes up to the point where I had to mash the reset button on my case to get out of it. Everything was fine till that point. 0o
edit: @nicstt: Well, what I meant by that was that the UPS was put in place to insulate my computer off from any voltage troubles coming from the electric company. Before I got a (previous, cheaper) UPS, during the rainy/thunderstorm season, fraction-of-a-second power dropouts would sometimes make my machine reboot. Having a UPS rules out voltage troubles coming from the electric compamy, and I don't really picture a fairly new UPS from a big-name company generating voltage spikes or sags, specially one of the more expensive, semi-pro ones. I later switched to this more pricier, beefier UPS because it turned out the cheaper one I was using before was rated for smaller-wattage computers. (It also turned out I was at the time using a lesser-wattage PSU than my machine called for, my hazey recollection is it might have been a 350 watt PSU, this was before I got the current, Corsair, 750 watt PSU.)
That could be a corrupt file then. Those hangs arer one step short of a shutdown. If you can roll back to an earlier restore point and get the update again.
Not sure if you checked this first but thought I would ask, are you rendering in 3Delight or Iray? I don't know about the Clothing Decals, but the link you're giving goes to WetFX Decal & Shader Kit, which as far as I can tell is Iray-only. If you're trying to render in 3Delight with it, then there's your problem. I use to have this happen to me quite a bit where I would be doing a render and within the next 1-5 minutes my laptop would completely shut down without any warning. If I recall correctly, this was due to trying to render scenes in 3Delight while using Iray shaders in it. Ever since I've avoided mixing the two, I haven't encountered any shutdowns during my renders.
Definitely not the issue. Daz automatically applies the correct uber shader when you try to render a scene with the incorrect shaders. I've run tests both ways, Iray to 3D amd 3D to Iray, and while the result may not look great it never crashes Daz much less shuts down the computer.
When ever a laptop hard resets the culprit has to first be assumed to be overheating.
I just about always use iRay. In any event, the main thing I was trying to find out is: do decals like the one I linked use appreciably more GPU horsepower than other, more mundane sorts of iRay shaders? I had done a couple of renders earlier of the same scene from more or less the same angle, the only difference between those and the render my machine suddenly shut down on is that I had then applied a couple different decals onto a garment, wasn't satisfied with one of those decals and so hid that one (turned it off in the scene via the eyeball symbol), set the viewport back to more mundane shader-based view so I could move around in the viewport smoothly again (because it was lagging like crazy in iray realtime view), and then I started a render with just the one decal remaining, so I could get a better idea how the wet splatter decal was going to look in a full render (instead of just an iray real time view of the garment) and a few brief moments into the render... FOOOM! the machine suddenly, completely shut off without warning.
So, just how extra render-horsepower IS a basic decal causing? Or DOES a decal even USE more rendering horespower than a mere iray shader?
The only things that can shut-down are hardware or driver issues. Though with the video driver windows should be able to recover. I'd try updating the drivers, all of them, even the chipset, and see what happens.
While rendering in Iray, your GPU has a finite amount of opperations per second that it can compute. If in fact your decals are using more GPU power, the end result will be a longer render time. An Iray render will peg your GPU at 100% usage at the fastest clock speed that it can while maintaining a safe opperating temperature.
Once again the only thing that will be affected by scene componenets while rendering is how long the render takes to complete.
i have seen a psu do this before ...normally its ok but heavy renders can draw too much power forcing the psu to power off automatically ...so maybe you need to upgrade psu which is easy.
That's not entirely true. Failures in Windows can be so severe you don't get the BSD but it skips right to a restart. I've seen it happen with software in development and sometimes just out of the blue for no clear reason.
Well, I pulled my machine out and dusted it inside a bit with compressed air can(s), checked at least superficiallly that the video card was still snugly in place... and connected the main cable from the PSU up to a PSU tester device.
I might be reading this wrong, but I noticed the -5V led was dark. 0o
I guess I need to order a new PSU from Amazon. 0O
Dang, what happened to the day when manufacturers made their electrical devices to last at least 20 or 30 years at a stretch? Now it's "Buy a new one in 5 years and throw the old one away!" **Facedesks**
Honestly 5 years is the time for throwing everything away and make a new pc .. Secially for 3d rendering where technology grows fast.
The amount of technology waste our generation is creating is honestly disgusting though
Well, the video card is much newer. I ordered that in August of 2017.
My lament isn't as much at computer makers, though frankly just from a quality standpoint they SHOULD make these things to last forever, even with the knowledge a lot of people won't want to keep them in service that long. But some will NEED to, at least for the imbedded applications. Having a machine that is running factory automation, say. My understanding is some of those sorts of installations are still running their machines off S-100 bus computers from the 1970s, simply because that's what existed when the machinery was originally installed, and to switch to more modern computers would require redesigning parts of the old factory machinery, and it just wasn't cost effective to change the rest of it just so they could switch to running it from a newer off-the-shelf desktop computer.
Awhile back I was watching a youtube video where this guy had this freestanding electric fan that was made in the 1960s, and was comparing it alongside a similar size and purpose electric fan from maybe 5 years ago. The 1960s one was all metal, and with a simple bit of cleaning and maybe a bit of oiling inside was as good as new. The much more recent one was largely plastic, and was already deteriorating and whatnot, with the fan blades wobbling and showing cracks. We need to get back to the era when everything was made to last 50 to 100 years, even the ephemeral stuff like box-fans and desk-lamps. I remember a TV ad for Curtis Mathis TVs, where you watched through jump-cuts an entire family generation go from the youngest being a little kid to him being an old geezer with grandkids, all sitting there watching the same TV. And then the TV quit working. The old geezer slapped his knee, laughed, and said "Well, I guess it's time to get another Curtis Mathis!"
Now, we've gotten well away from the era of quality-for-quality's-sake, and into the era of "Lets make a fast buck, nothing lasts forever! They'll buy another eeeelectric fan in 5 years anyway..." It shows a complete LACK of integrety in our big business world.
There is no -5Volt rail on ATX PSU's. There's a 3.3, 5 and 12. According to that tester your PSU's voltages are right.
It's what people want. You can still find appliances and other things that are made of metal, and will last for generations. But they're expensive. I have a five-year-old Sunbeam that is made of metal. I picked it up at a yard sale for five dollars, the woman was selling it because she'd remodeled her kitchen and it no longer matched the decor.
People are like that, they want cheap throwaway junk, and the companies make it for them to buy. It's not always a matter of corporate integrity, it's often a matter of giving people what they want.
Well personally I'm a dedicated collector. I even have a working Playstation one with a large set of original games. Not to mention all sort of toys and comics. So I guess it's also a bit of your choice what to do. I mean, I still have fun with that old Playstation, but I can't render iray with it ..
Unfortunately, those types of power supply testers are not definitive when it comes to making decisions on the health (or lack thereof) of a PSU. They are better than nothing when it comes to determining whether the PSU is working or not, but to truly test the PSU you really need a digital multimeter than can display the actual voltage outputs on the various rails. It is also important to do so not simply when the computer is idling, but also when under load. The results don't need to match the nominal voltages exactly, but should be within 5% to be considered acceptable.(10% for the -12). If you don't have access to a digital multimeter and don't wish to purchase one, perhaps ask around, as a friend might have one that you could borrow - they have more uses than just computer electronics.