Holy cow, we're back! >>18190
Since we don't have a /g/ or a /tech/ on the webring...
I've gotten some new games and tried some new demos, but after playing them for a bit (even sitting in main menu) my screen goes black and becomes unresponsive. It's not just the game, but the entire system. My USB devices are still getting power (though my Dualshock 4 gets a red light when I have it set to no light) and my fans are still going. My display indicates it has a signal, but can't tell me a resolution.
I'm running Windows 7. No issues seem to happen with older games (or new re-releases of old games in the case of Red Faction Guerrilla Re-Mars-tered), but I don't have any old games that are also demanding on my PC handy. I managed to go a decent distance in Travis Strikes Again with it in windowed mode, but it happened within 2 minutes of me setting it back to borderless window mode. After rebooting my fans go wild for a few seconds once I hit windows login (they don't even get particularlly active before this starts), then go back to normal.
I've tried both the most current drivers for my graphics card (RX480) and the oldest ones AMD has, but both have the issue. I've tried it on a few games, so it's not an engine wide issue. I've had GPUs fail on me, and it's not consistent with those times either.
Any ideas on what is causing this issue?
How are your temps? It sounds to me like your gpu's cooking itself. Use Afterburnet to check for temps mid game.
My thoughts exactly. When was the last time you changed thermal paste on your GPU?
Are you running a dual screen set-up? If so, you could use HWinfo or some other monitoring software to see the temperatures.
>thermal paste on your GPU
what do you mean "?"
Do you not know there's a thermally conductive paste between the GPU and CPU chips and their respective fans/cooling systems? Take the GPU out and take off that plastic casing, look for any signs of burn damage.
In case you really are that new to PCs, gif related is a joke, you do not put mayo in the CPU slot.
I put thermal paste on my CPU. I've just never heard of putting it on a GPU.
You've heard about it now. Get your phillips head screwdriver and start unscrewing.
That almost sound like a botched power surge protection trigger.
Which given the RX480 might actually be what is happening, is it a 3rd party model with 8 / 6+8 / 8+8 pins for power or a stock AMD one with a single 6pin for power?
Is your mobo from ASUS and is it a gen 1-6 intel core platform one?
What's your PSU? (wattage and brand at the very least but exact model isn't unhelfpul)
What drivers? Tried DDU and newest / stable older drivers yet?
There's thermal "paste" between any heatsink and whatever is sinking heat.
I know there's paste in it, but never heard of user replacing it.
>is it a 3rd party model with 8 / 6+8 / 8+8 pins for power or a stock AMD one with a single 6pin for power?
I'll have to pull that up.
Gigabyte for i7-3770k
>What drivers? Tried DDU and newest / stable older drivers yet?
Already mentioned in OP
That may actually be it. It's the only thing I've changed since this started. I'll have to swap to the old one and test when I get a free day.
APC UPS BE850M2, 850VA
>but never heard of user replacing it.
Because you're a newfag. Even the most expensive best-rated thermals GPU has trash thermal paste compared to what you can buy. Enthusiasts always use their own bought paste.
>Already mentioned in OP
My bad, read too fast
>I'll have to pull that up.
Well if you do have only a single 6 pin, do not ever install the RX480 launch drivers and try to run anything intensive, you can fry your mobo if you do.
>Gigabyte for i7-3770k
Highly oc'ed or stock / near stock?
>I know there's paste in it, but never heard of user replacing it.
Pretty much every manufacturer uses trash that dries in a couple years and is actually bad for your cooling after that point, not that it was even that great to begin with.
If you do replace it though be warned that on a GPU you're dealing with a naked die not an IHS like most CPUs in the past 15 years so be careful and avoid anything conductive unless you really know what you're doing.
If you're not too good with applying thermal paste get the Artic Silver Céramique (or Céramique 2) paste, it's not the greatest ever and it dries over time too (although it takes a while before it's really noticeable) but it's super cheap, fairly good and basically the easiest to apply paste you can find
>APC UPS BE850M2, 850VA
That's an UPS not a PSU, PSU is the thing inside the case
Although that also adds the possibility that your UPS is in need of a new battery or you're going over capacity on it under certain load scenarios.
Do you have the monitoring software for that thing installed and if so does the logs for it say anything, also check the battery replacement warning / last replacement date thing, you don't wanna use those batteries for over 5 years
Can you also check how high the load goes when you're doing something that causes the issue?
If you don't have it
UPS is only a couple of months old. Thanks for the warning on the oldest RX480 drivers. I'll try a middle upgrade.
Any recs for good monitoring software for GPU load?
We alrady mentioned them in this thread:
GPU-Z also has a GPU load test integrated so I prefer it over the other
unlikely but I have heard of issues with newer games running ryzen/windows 7 on unsupported chipsets
>Any ideas on what is causing this issue?
- Overheating (CPU or GPU)
- GPU about to croak
- unlikely, but could be a software issue (corrupt Windows files or some software running in background conflicting with your games, i.e. TeamViewer).
- unlikely, but could also be a bad HDD/SSD.
Here's the software you should start with:
1. MSI Afterburner (to monitor GPU temps and GPU load)
2. Core Temp (to monitor CPU temps and CPU load)
3. Memtest86+ (test RAM for errors) *
4. OCCT and Furmark (test GPU for errors) **
5. CrystalDiskInfo (check SMART data for errors) - however if you can I'd recommend installing a pirated copy of HD Tune Pro or HD Sentinel Pro since their SMART info is more complete and accurate.
* Make sure you get Memtest86+ and not Memtest86, it's a bootable program so you'll need it to boot from a USB thumb drive or CD
** Be sure to get Furmark ASUS ROG Edition since it's the only one with GPU artifact testing functionality
Also OP, if you're going to roll up your sleeves to change thermal paste on the CPU/GPU and/or test another PSU, also have a pencil eraser and contact cleaner spray or isopropyl alcohol in hand.
Clean the contacts on your ram and GPU with the eraser until their color changes from bright gold or brown or red depending how badly oxydized they are to almost a bright white, then rinse with contact cleaner spray or isopropyl alcohol.
And for God's sake, wait for it all to dry before sticking it back in.
Here's my log from GPU-Z while running a game and having this happen. Any obvious issue from this?
I also have a correction. My display is showing no signal at all and not all USB devices maintain power (I forgot my network stick has a light on it because it's normally covered by something else)
>Any obvious issue from this?
Nothing obviously wrong here, power draw is mostly what you would expect temps aren't outrageous, voltages are fine
And don't forget to clean both sides lol
It's a 200 KB text log in barely readable formatting. I'm not going to wade through that shit or try parsing it in a spreadsheet trying to find anything unusual unless I was getting paid to do it. You should at least edit the log to only show the relevant content if you noticed anything unusual. Or make a screenshot of the GPU-Z graphs instead.
I've noticed that in these situations it's not unusual for USB devices to act up.
> in barely readable formatting
Open it in a text editor instead of your browser.
Turn word wrap off .
Thanks, just realized that.
I didn't notice anything unusual either but temps are a bit on the high side. However notice that while the GPU reaches 75 celsius, the fans never go over 30% of max speed. In almost all video cards, default BIOS fan profiles tend to be too liberal with temperatures (they let them get too hot before kicking in I suspect it's a jewish ploy by manufacturers to make cards have a shorter life than they should). You can create a custom profile using MSI Afterburner (however it only works while the program is running in the background).
After I created a custom profile, my MSI GTX 1070 never gets hotter than 60 celsius with full load in hot summer days, and that's with the fans never going more than 50% speed. With the default fan profile the card would often go over 80c, and the fans don't even move before it gets to 70c.
Is my soon to be 8 years old i7 3770K enough for another year or so? I hope it is because a bunch of games are targeting both current and next gen and my GPU+CPU have still been running PC ports decently.
Maybe a year at best. Don't expect to be doing well with the multiplats made for next gen consoles. I'm already feeling like it's way past time to replace my 4770k and I'm going nuts waiting for Zen 3 to come out. The wait is a bitch.
Unless you're emulating modern systems what even is going to push your CPU?
>inb4 unoptimized eternal early access trash
Just checked my old APC. Not APC related.
A year or so is all I want. Other than Serious Sam 4 and Cyberpunk, I don't think the other upcoming games on my radar will be too much.
It should get you by this year but AMD has some good offers for CPUs that you could jump on considering you find a fitting board too.
I have a RX580 and had a very similar issue to OP. Whenever I ran games it would random turn the whole PC off. I replaced every part until I solved the problem.. Asus motherboard was causing it. Something about my set up and the motherboard hated each other and it wasn't just a single Asus board, I tried 2. Ended up switching to a tomahawk and it's worked like a dream since.
Get a copy of Darksiders and try running it. For my issue I found Darksiders would instantly trigger the shut off. Within 5 seconds of gameplay it died. It's on GoG so it's easy to pirate if you don't own it.
Have it. I'll try it when I get some time.
Two new logs. First is Darksiders menu which didn't fail. Second is some Light's Challenge in Mega Man 11, as that's the last game I played without this issue. MM11 has no issues that I see.
Didn't try gameplay because I went through the menu without the main issue, but frame rate remained uncapped even after turning on v-sync so I'm not playing around with it right now. Go to ask: How long did you use your setup before the issues started? My machine has served me for years, but you imply it was there at the start.
Mine was immediate. If yours is old and fucking up just replace it
Thing is I have no idea if it's a mother board issue.
Did you run a full Memtest86+ pass, did you run OCCT and Furmark gpu tests, and loaded CrystalDiskInfo like I suggested you do? What were the results?
I'm waiting on some discs to burn memtest. Should be here soon.
Did OCCT in the meantime. ~Halfway through the GUI disappears but it kept using resources till I killed it via task manager several minutes later. This (which I had to rename to .txt to upload) was in the directory I had the EXE and the modified time indicates it appeared when the GUI disappeared.
Memtest also supports USB drives if you have any USB thumb drive around.
How long did it take after you started OCCT until the GUI disappeared? Did you see if it detected any errors before it crashed?
OCCT also has a PSU test (the "Power" test in main menu). You could try it and see if it crashes faster.
Roughly halfway through. Didn't see any errors displayed.
>roughly halfway through
Which is...? The test has no time limit or progress bar / percentage, it keeps running for as long as you allow it.
Realized I forgot to mention that when I was running the power test, but didn't want to interupt it. I set the general test to 30 mins (the default for non-indefinite), so ~15.
Just finished 30 mins of the power test. No errors.
allow my pajeet-tier tech support skills to guide you
it's possible that you may have gotten two fucked drivers in a row. did you only get the latest and oldest ones available? the latest ones are always bloated, unpatched pieces of shit, and the first version usually is no better with tons of launch day bugs. highly unlikely, but try getting a driver that's in between. I'd go for one that's a couple months old and the latest iteration of it's version (for example if it goes from 3.6.xxx to 3.7.xxx, get the most recent version of 3.6)
it'd also be a good idea to check that Wangblows isn't fucked, run these in a command prompt as admin:
that checks for corrupt windows files and bad drive sectors, respectively. running a full drive health test would also be helpful (WD Data Lifeguard Diagnostics is a good one)
it could also be a hardware issue, which usually causes more rampant and unpredictable issues such as artifacting and random crashes with different errors each time. if you have a spare GPU, slap it in there and play your gayms for a while (make sure you get rid of any old drivers/software as well). if the issues persist, it's a separate issue not related to your GPU.
scan for adware/malware as well, check what programs are installed and what starts up with your machine, uninstall what you don't need (use a real program like Revo, not the built-in windows one that leaves shit behind)
I've tried three drivers version now. I'm currently on a middle version and still have the issue.
Did a malware scan with malware bytes. Are there better free programs these days?
Any idea how long sfc /scannow and chkdsk take so I know how much time I'll need to set aside? Current storage is a 120GB solid state+3TB HDD+2TB HDD
Note I'll try the general scan when I can leave my computer running at max for 30 mins and keep an eye on it. See if that was an abnormality.
Good. I'd say this result reduces the likelyhood of a PSU issue since that test pushes both CPU and GPU to the limits.
Now I'd try running the 3D test again and see if it crashes OCCT again. And then I'd run Furmark's artifact scan test.
Good advice on commands, though normally you run chkdsk /f first and then sfc /scannow.
I'll add that one last resort if all else fails is a BIOS CMOS reset (though one can try it earlier as well if they're familiar with BIOS settings).
chkdsk /f (which doesn't do a full surface scan, just checks the file system) is usually quick unless your drives are very full and fragmented, shouldn't take longer than 15 minutes in a worst case scenario.
For a full surface scan on all these drives, you should start it before going to bed before a day you're going to spend outside, with luck it'll be over by the time you return.
sfc /scannow also shouldn't take longer than 15 minutes.
malwarebytes is probably the best you're going to get, if it doesn't catch anything nothing else will
sfc and chkdsk don't take too long, maybe 10 minutes total.
forgot to mention, a BIOS update would help too, unless you have an intel board then you're fucked since they nuked all of their motherboard drivers/BIO files from their site
I had to update my bios to get the current graphics card to boot (wouldn't even boot to bios on the stock version). I'd need an entirely seperate and older graphics card on standby (which I don't have) to even navigate BIOS if I started fucking around in there. I really doubt there's a newer version since the one I installed was the only one offered, and it was years old. I'm of the position I'd just build my new PC a few CPU gens earlier than I intended than do anything with the mobo.
I'll run the tests when I get some time (likely this weekend). CD-R should have come by then too.
I have a Micro ATX case with really shit cooling, whenever I play vidya my GPU stays around 80C. problem is since it's so small there's not much I can really do to increase airflow. GPU also sags a bit so I have a riser attached to it from the slots on the back, which further restricts airflow from the front side of it if you're looking at it from the side. the best thing I could think of would be to drill some holes in the bottom of it and get some taller feet so there's a sort of exhaust from the bottom. Any other ideas?
Where the fuck is that front fan even pulling air from? Drilling holes in that front panel in front of the fan would probably increase air flow significantly. I hate that all these retarded cases without mesh front panels exist.
there's some small gaps under the front bezel where it gets air. the biggest issue isn't getting cool air in, it's getting hot air out
>Any other ideas?
Use MSI Afterburner or GPU-Z to check how hard your GPU fans are working when your GPU is running too hot. Chances are they're not even at 50%. If that's the case, use MSI Afterburner to create a custom GPU fan profile >>13269
When it comes to physical workarounds your ideas are probably your best options, other than getting a better case (or doing the crude but effective method of keeping the case open with a table fan blowing directly on the components).
Yeah, the GPU is basically only getting air from whatever it can pull from your side panel perforations, and the small amount of air that fan is pull from the bottom of that panel. If you don't want to fuck up the front panel you could try mounting the fan to that side panel instead as intake. I'll admit I'm no expert on air flow and computer configurations and am just going off of my gut and the slight information I've gathered researching my own future builds. But I'm sure if you test various fan placements with benchmarks it shouldn't be too hard to see if it's bringing your temps down. What a fucking horribly designed case.
are there any libre note-taking applications similar to OneNote? I need one that lets me sync to a central device (my home server)
No idea about the act of note taking in and of itself but syncthing can p2p files very easily and comofrtable while not caring about a specific program
yeah I was just thinking about using something like FreeFileSync and just use that to sync everything to my main server. I was deadset on using NextCloud but I'm starting to find better replacements for it instead of setting up a webserver and paying for a domain (which you have to use apparently if you want to access it from outside your network)
Both report they found no problems.
My CD-Rs came, but it's a little late for me to start memtest. Any idea how long a memtest86+ takes?
Any interest in a /tech/ on Zchan?
>Any idea how long a memtest86+ takes?
A single loop isn't very long but ideally you just want to run it for a few hours just to be sure.
You can also use the RAM tester that's integrated into Win7+ by going into the boot menu (F8 when windows is booting).
I had a similar issue with my GPU back in the day. It started happening when my chink bought PSU blew up. Even after I replaced it with a silver standard PSU the GPU would hit dangerous temps even on basic indie pixelshit. I ended up replacing the GPU since it was just a shitty 660 but I assume some circuitry got fried in it due to the PSU blowing up.
We now have a >>>/tech/.
Found your problem anon. Not the hardware itself, but Radeon drivers have been fucked since they released the 2020 update at the beginning of December last year.
The most recent couple of updates have been geared towards fixing, or trying to fix the problems.
They spent the first few months denying the issue existed
If you have anything open in the background that has hardware acceleration on, switching it off usually mitigates some of the problems.
The last driver release that most people agree was stable was 19.11.3 but it's hit and miss for a lot of people, same with games.
I thought it might be that, but it happened with launch drivers too. I'll try the speciffic drivers you listed though.
>Found your problem anon. Not the hardware itself, but Radeon drivers have been fucked since they released the 2020 update at the beginning of December last year.
Fixed as of the most recent drivers I used but importantly you needed to install the one they labelled as optional/experimental.
To anyone who has Steam running through Wine these days, how the fuck did you do it? It's been years since I've installed this shit used winetricks this time, -no-cef-sandbox does not make the library or other browser pages visible, and the Windows XP workaround doesn't work now that Valve has dropped support for it.
I just want to play Pathologic.
Just use Lutris. It's much more convenient than managing the headache of all those separate installs and libraries yourself.
Nope, even happens on 19.11.3.
I think I'll do that, thanks. The Pathologic install script looks similar to the instructions on Wine's AppDB but without the added pain of setting up an extra Steam installation.
Welcome to webshit, Steam is now 99% Chromium and HTML/CSS/JS. You can still view the library without CEF by passing the starting URL of steam://open/minigameslist to Steam on startup, so something like "steam steam://open/minigameslist".
Yeah, it helps that the Lutris install script is generally the best/most compatible possible way to set up the game. when it's available. There may be some edge cases but I'd be hard pressed to find them.
Just pirate the GOG version ffs
Still didn't work and I doubt downloading the GOG version would make any difference. Maybe I'll just play Pathologic 2 in the meantime and try the older one when I have a Windows install again.
Maybe I'm competely retarded but couldn't you just use the linux version instead of trying to install it with Wine?
I don't know a thing about Pathologic specifically, but if they released it natively for Linux back in the day then it is likely dependent on libraries that are not available in modern distros. Modern Linux ports have a cross-platform API buffer (SDL, Vulkan) for most of the libraries that make it easier to deal with the constantly changing state of Linux libraries.
>I doubt downloading the GOG version would make any difference.
Unless I have missed something, you can't play Pathologic because you can't get the Steam client running.
If you didn't realize that on your own, GOG version doesn't require Steam.
Just gave memtest86+ 5 hours. Made 4 passes with no errors found.
Oh, I sort of got Steam working through Lutris but Pathologic itself still didn't work.
Are you joking? Steam is in virtually every single linux distros repos. Steam is a native linux program.
Steam on Linux just runs non-compatible games through their own branch of Wine, called Proton. It's pretty decent, but usually a Lutris script will use that version if it's better, and not if not. Again, sometimes if the game is recent or there was a recent discovery made in compatibility, it might not be up to date. But Pathologic is pretty old.
>the absolute state of loonixfags
Of course, they'll also call you a shill and claim that Linux runs 110% of windows games at better performance than native in the next thread.
Waste of trips
Good job anon. Did you run CrystalDiskInfo or other program to check SMART data on your storage devices? Everything's fine on that end?
If that's the case, then I guess it's time to roll up your sleeves and mess with hardware.
Clean the contacts on your RAM sticks and GPU as described in >>13166. See if that has any effect fixing the issue.
If that yields no results, the next best step to do would be resetting BIOS settings back to defaults. I understand you have problems even getting into the BIOS, but it's really the best option - unless you'd prefer to try reformatting and reinstalling Windows first or keep messing with different driver versions.
Keep in mind that resetting the BIOS will NOT revert it back to an old version (since you're worried that you had to update the BIOS to get your gpu working). It will only reset settings and PnP hardware data. Your motherboard has a VGA or HDMI connector, right? What you should do:
>1. remove your video card (ofc remember to unplug the PSU power cable first)
>2. reset the BIOS using the CLR_CMOS jumper on the motherboard (refer to your motherboard's manual for its location)
(if you don't have a jumper piece available you can use any screwdriver, just short the jumper pins with its tip)
>3. plug your monitor into the motherboard integrated GPU connector
>4. boot the PC, get into the BIOS and set it up - configure the BIOS to boot through the PCIe card first
>5. power off and unplug, put the video card in (don't forget about the PCI-E power connectors)
Oh, I thought you were talking about Pathologic.
It runs almost all of my vidya either natively or very well. Pathologic is an exception because it's unholy slav shitcode written by people who'd never made a game before. It may actually be the only game I have that I can't run in Wine and even then others have managed.
Looking into the AMD black screen issue a bit more suggests I may need to use an external program to clear traces of old drivers from my PC (even though I picked clean install on driver install for the stable version). Thoughts on this and what program to use for it?
DDU (Display Driver Uninstaller) - which was already mentioned in this thread, and you even replied to and fucking QUOTED the anon who mentioned the program asking if you had already used it. Try being less of a dislexic sperg and paying a little more attention to anons trying to help you, ffs.
I assumed you had already used it, others probably thought the same.
DDUed drivers and went to 19.11.3. Still happens.
I don't have integrated graphics.
UE4 games are mostly fucked for me on AMD. The Chinese must be to blame.