1050 ti acting strange, possibly dying but too young to die. HELP

[PotatoPCGamer]

not sure if it's a problem with the GPU or the OS but I'll give it a try

So this morning the computer works fine, I was playing gta iv when suddenly a BsoD pops out. it was because of the video_tdr_failure or something like that, which says about nvlddmkm getting problems. after 5 hours of tampering, trials, and errors (also there's a few-seconds power outage somewhere in the middle of that 5 hours) now everytime I tried to game after 5-30 minutes BsoD would appear, the last one before this text is about video_scheduler_internal_error.
I'm really tired of trying everything the internet has to offer. That includes : virus checking (no result), registry cleaning, SFC (said my pc was fine), DISM, CCleaner, reinstalling drivers and stuff, furmark benching just to spot artifacts (none), retesting with some games, wiggling cables and the GPU itself inside my pc, and some more. Now it's night and I as a human needs to get some rest, but fortunately the system wants to joke around.

I have grown tired of trying. The GPU is just over 2 weeks old, and the rig itself too. In my place it's not easy to claim warranty of this gpu. The only things I haven't try are : power off and unplugging the main powering cable, fresh install windows, and bringing the pc to mechanics.
the last option isn't that preferable since everyone is basically indian tech support scammer with more homelandish look.

I'm no technician but I can tell this is either the GPU or the OS. I need help quick, I don't want my money to vanish just like that, in here a 1050 ti costs around $170 converted, which is bs.

I need professional help, but don't expect me to understand some nerdy programming stuff regarding the pc. Thanks for reading this long anyways

I never clock my rig, all parts of it

TL;DR : video_scheduler_internal_error, pc got possibly rekt, plz help

Sorry for the possibly uncomplete English, it's not my main language.

specs :
Win 10 64 Bit
2x4Gb Corsair RAM
Inno3D GTX 1050 Ti Single Fan
I5 4440
Mobo GA-H81M-S2PH
1TB Seagate Barracuda SATA 2
Dell U2412M 24 inch 1920x1200
PSU Bigfoot H500G

HEELLPPP!

EDIT : I have seen an anomaly which is when Furmark is launched and runned it would say the GPU is 47 degrees hot after 5 seconds, but when I open MSI afterburner STRAIGHT AFTER the furmark it still says 40 degrees. I have no more time to deal with this so I'll leave this for the night + school (around 18 hours)

EDIT 2 : So I have been noticing, every single time before the crash when I'm gaming there's a load of purple/red/white squares popping out, and then not a second after that it freezes, sometimes with a seizure-inducing flash of color boxes before a BsoD. What does this mean?

 

[johnbl]

Most for the time when I see a video time out on a Nvidia card it is not caused by the card.
As a general fix, you would update the bios, update the motherboard drivers from the motherboard vendors website.

Bad/old motherboard sound drivers screw with the gpu hdmi sound support and cause delays that result in a timeout.
Nvidia also enables shadowplay by default, old network drivers introduce errors that slows the shadowplay and results I. Timeouts. I would disable shadowplay and sound sources that do not have a speaker attached.
Timeouts can only be debugged with a kernel dump, but any memory dump can help isolate the issue. You might post link to your memory dump.

Overclock software can also cause this error. People overclock the card and might need to increase the pci/e bus speed so the electronic Signals have the correct timing windows. Default for a pci bus clock will be 100 MHz in bios.

Changing video cards without going into bios and resetting it to defaults can also cause timeouts. It can force windows plug and play to use really bad hardware settings.

 

[PotatoPCGamer]

Most for the time when I see a video time out on a Nvidia card it is not caused by the card.
As a general fix, you would update the bios, update the motherboard drivers from the motherboard vendors website.

Bad/old motherboard sound drivers screw with the gpu hdmi sound support and cause delays that result in a timeout.
Nvidia also enables shadowplay by default, old network drivers introduce errors that slows the shadowplay and results I. Timeouts. I would disable shadowplay and sound sources that do not have a speaker attached.
Timeouts can only be debugged with a kernel dump, but any memory dump can help isolate the issue. You might post link to your memory dump.

Overclock software can also cause this error. People overclock the card and might need to increase the pci/e bus speed so the electronic Signals have the correct timing windows. Default for a pci bus clock will be 100 MHz in bios.

Changing video cards without going into bios and resetting it to defaults can also cause timeouts. It can force windows plug and play to use really bad hardware settings.

How do you upload memory dump files here?
EDIT : nvm, doing it on mediafire.

Minidump files : http://www.mediafire.com/file/hl6dpo6oue74vw4/032717-22500-01.rar
the rar name is also the name of the most recent dmp file.

MEMORY.DMP :
http://www.mediafire.com/file/3pmpnx4gzjgh04x/MEMORY.DMP.rar

Also, there's something bugging me : everytime a game crashes there would be white/purple squares popping out, and then the crash, sometimes with a display full of white/purple/red color boxes across the screen. does this mean anything?

And if I change to another OS would the problem still occur? I heard people ranting about win 10 crashing games but mostly it's on 2016 or less

 

[RobCrezz]

Could be that PSU giving bad power delivery to the GPU. Ive not heard of that brand, could be bad..

 

[PotatoPCGamer]

Could be that PSU giving bad power delivery to the GPU. Ive not heard of that brand, could be bad..

how do you test the psu power delivery via desktop? the cables are already arranged so neatly that it's a hassle and a waste to pull the off :3

no, seriously, I don't have a backup psu.

 

[helpstar]

try the card in another pc

Also power outages or unstable voltage can be possible to cause this problem

BIOS updated to version F7 via q-flash? Don´t flash in windows, use the BIOS integrated tool q-flash
http://www.gigabyte.us/Motherboard/GA-H81M-S2PH-rev-10#ov

use "DDU uninstaller" and delete all old drivers, AMD/nvidia and after this, install the latest driver from nvidia.com

 

[PotatoPCGamer]

try the card in another pc

Also power outages or unstable voltage can be possible to cause this problem

BIOS updated to version F7 via q-flash? Don´t flash in windows, use the BIOS integrated tool q-flash
http://www.gigabyte.us/Motherboard/GA-H81M-S2PH-rev-10#ov

use "DDU uninstaller" and delete all old drivers, AMD/nvidia and after this, install the latest driver from nvidia.com

I don't have another pc, the only otehr thing I have is a laptop.

I tried using ddu and fresh installing the driver, worked for a night, but then the crash occurs everytime I'm quitting a game or any app , even youtube if i'm streaming on 1080p or above. Thankfully not when I'm gaming itself, it have to be when I'm quitting the application.

Power is stable here, and I'm using a volt stabilizer.

About the bios flashing, some said that any problems occur when flashing could wreck my pc, I'm a bit scared to do that now, maybe later.

I'm looking to try to install another os like win 8.1 or 7, but looking for an original one is really hard here, I'm not allowed to use visa since I'm just below 17, and pirating seems wrong.

 

[helpstar]

So the PC has problems while getting from load to idle, could indicate a faulty PSU or volt stabilizer.

Your motherboard has a dual BIOS, which means there are two chips physically on it. If something wents wrong with the chip you wanna update, the backup BIOS automatically flashes the primary BIOS back to the old version, so there is no risk at all. Just flash with the BIOS integrated flash tool. Don´t use windows, to minimize any failing while updating.

You could try a windows in "demo" mode (testversion), should be running like 90 days without needing to activate.

Maybe your PSU doesn´t support C6/C7 C-states (haswell ready). Disable them in BIOS.

 

[PotatoPCGamer]

So the PC has problems while getting from load to idle, could indicate a faulty PSU or volt stabilizer.

Your motherboard has a dual BIOS, which means there are two chips physically on it. If something wents wrong with the chip you wanna update, the backup BIOS automatically flashes the primary BIOS back to the old version, so there is no risk at all. Just flash with the BIOS integrated flash tool. Don´t use windows, to minimize any failing while updating.

You could try a windows in "demo" mode (testversion), should be running like 90 days without needing to activate.

Maybe your PSU doesn´t support C6/C7 C-states (haswell ready). Disable them in BIOS.

hmm, it seems disabling c6/7 does the job for now (also disabling c3 just in case), combined with fresh driver install.
let's see for a few days tho, if somehow it came back, or if luckily it dont

 

[PotatoPCGamer]

So the PC has problems while getting from load to idle, could indicate a faulty PSU or volt stabilizer.

Your motherboard has a dual BIOS, which means there are two chips physically on it. If something wents wrong with the chip you wanna update, the backup BIOS automatically flashes the primary BIOS back to the old version, so there is no risk at all. Just flash with the BIOS integrated flash tool. Don´t use windows, to minimize any failing while updating.

You could try a windows in "demo" mode (testversion), should be running like 90 days without needing to activate.

Maybe your PSU doesn´t support C6/C7 C-states (haswell ready). Disable them in BIOS.

Disabling C6/7 and also C3 seems to do the job, no more crashes until now, hopefully also for the future. thanks

 

[helpstar]

C3 (deep sleep) shouldn´t be a problem.

Sounds like the PSU is just not compatible to the "new" C-states of CPUs since the Haswell generation.

 

[johnbl]

- looks like a pirate version of windows running 2 copies of a key server

- Ethernet driver is having some issues also, you should update it
(Realtek NICDRV 8169 PCIe GBE Family Controller driver)

- you have a old DVD driver from 2011, likely to be having problems. I would just remove it.

you have two audio devices disabled, you can re enable them if you know they are not your problem.

 

[PotatoPCGamer]

- looks like a pirate version of windows running 2 copies of a key server

- Ethernet driver is having some issues also, you should update it
(Realtek NICDRV 8169 PCIe GBE Family Controller driver)

- you have a old DVD driver from 2011, likely to be having problems. I would just remove it.

you have two audio devices disabled, you can re enable them if you know they are not your problem.

I have changed alot since the time I published the dump files, but thanks anyways.
dammit never known those "real" vendors offering "real" everything gives pirate windows. I can't really complain tho since most of this rig were original, but maybe after I got more money to buy the real one I'll buy one, or I'll just switch to linux or something, to solve the problem.

which dvd driver? can't find it since everything is so jumbled. I have updated the ethernet driver, and those 2 audio devices aren't the issue but not needed anyways.

 

[PotatoPCGamer]

spoke too soon, it came back.
purple/red/white boxes flashing here and there, and after about some miliseconds bluescreen occurs.
anyone know how to fix this? this time, fresh driver installing didn't work

 

[surya13]

i feel that your psu is the culprit. do you have the problem only while playing gta 4 or while playing all games??

 

[PotatoPCGamer]

i feel that your psu is the culprit. do you have the problem only while playing gta 4 or while playing all games??

all games. CSGO, GTA IV, BF1, and else all crashed as soon as I'm loading in from the main menu to the actual game.

 

[surya13]

Install MSI Afterburner and reduce your card's power limit to 80%. Play a game at the lowest possible settings. The aim is to put less load on the GPU and hence PSU. See if your issue persists. If the issue persists, reduce the power limit in steps of 10% till the least possible and see what happens.

 

[PotatoPCGamer]

Install MSI Afterburner and reduce your card's power limit to 80%. Play a game at the lowest possible settings. The aim is to put less load on the GPU and hence PSU. See if your issue persists. If the issue persists, reduce the power limit in steps of 10% till the least possible and see what happens.

no difference, even with minimal everything and maximal fan speed it still crash every 10 minutes, and a split second before the crash red/purple stars would pop like crazy.

btw, I tried my friend's windows 8.1, same problem.

the game (War Thunder) crashed, and when I alt-tab something like this showed up. any meaning of this? after I alt-tabbed into the game again and somehow able to manually close the game via itself my desktop went back to normal

 

[surya13]

Set the power limit to the least possible. Then try some games old like Need for Speed Most Wanted 2005 in maximum settings with vsync disabled.

 

[PotatoPCGamer]

Set the power limit to the least possible. Then try some games old like Need for Speed Most Wanted 2005 in maximum settings with vsync disabled.

not that I don't want to, but the sole purpose of buying the pc is for nowadays gaming >,>.
I'll try half life 2 tho, I have the game legit on steam.

also, i finally have time to run memtest, and the result is none, everything is working perfectly.

EDIT : Half life 2 lowest : 300 fps avg no crashes nothing
Half life 2 medium : 300 fps no crashes but a sometimes a random white polygonal shape would flash onto the screen

EDIT 2 : I have been observing, after I played the hl2 war thunder didn't crash anymore, still the purple red stars appeared every once in a while and some stutters are going.

EDIT 2.5 : after I restart the pc everything came back, all the crashes and purple-red stars and boxes.

EDIT 3 : Uploaded my DxDiag, just if someone could check it if my pc had something wrong (win 8.1 one, not win 10) : https://pastebin.com/udR3x4Ud

EDIT 4 : Battlefield 1's crash report looked like this :

 

[surya13]

I feel that it is too early for a GPU to show signs of dying. Since you don't have another GPU or PSU, it is gonna be difficult to isolate the problem and identify the culprit.

I guess that you have this version of 1050Ti
https://www.primeabgb.com/online-price-reviews-india/inno3d-gtx-1050-ti-compact-gddr5-graphic-card/

It doesn't have a power connector and draws power from the motherboard. 500W PSU means that it is capable of delivering 500W power with all its cables connected. Your motherboard may take around 150-200W under load (CPU+GPU+RAM+others). Your PSU is not of high quality and may not be able to supply that required wattage.

How long is your PC able to run furmark before crashing?

Try to put load on your PSU by pushing your processor to its limits. Run the latest version prime95 torture test. Run the test in all the available torture test modes for at least 2 hours each and see whether your system is stable or not.

Edit: See which setting of prime95 consumes more power and use it. Here our aim is not to test the stability of the processor but to test the PSU. So it would be more relevant to use the most power consuming option. Use IntelXTU to monitor CPU power usage (enable Package TDP under monitoring). Try to run the test for at least 2 hours or till the system crashes.

Edit 2: I have had a similar issue with my old GT610. I had a low quality 350W PSU. While playing games, textures will get replaced by random colors and the game will eventually crash. Sometimes the system itself will freeze and wouldn't respond. If I try to restart, it will not turn on for at least 10 minutes after the system crash. I have also got BSODs. The PSU blew up after a few days and I replaced it with a 500W PSU (low quality one again. Who bothers? It was a 6 year old PC at that time) and it worked fine after that.

 

[helpstar]

Disconnect and reconnect the Card from the mainboard slot

 

[PotatoPCGamer]

I feel that it is too early for a GPU to show signs of dying. Since you don't have another GPU or PSU, it is gonna be difficult to isolate the problem and identify the culprit.

I guess that you have this version of 1050Ti
https://www.primeabgb.com/online-price-reviews-india/inno3d-gtx-1050-ti-compact-gddr5-graphic-card/

It doesn't have a power connector and draws power from the motherboard. 500W PSU means that it is capable of delivering 500W power with all its cables connected. Your motherboard may take around 150-200W under load (CPU+GPU+RAM+others). Your PSU is not of high quality and may not be able to supply that required wattage.

How long is your PC able to run furmark before crashing?

Try to put load on your PSU by pushing your processor to its limits. Run the latest version prime95 torture test. Run the test in all the available torture test modes for at least 2 hours each and see whether your system is stable or not.

Edit: See which setting of prime95 consumes more power and use it. Here our aim is not to test the stability of the processor but to test the PSU. So it would be more relevant to use the most power consuming option. Use IntelXTU to monitor CPU power usage (enable Package TDP under monitoring). Try to run the test for at least 2 hours or till the system crashes.

Edit 2: I have had a similar issue with my old GT610. I had a low quality 350W PSU. While playing games, textures will get replaced by random colors and the game will eventually crash. Sometimes the system itself will freeze and wouldn't respond. If I try to restart, it will not turn on for at least 10 minutes after the system crash. I have also got BSODs. The PSU blew up after a few days and I replaced it with a 500W PSU (low quality one again. Who bothers? It was a 6 year old PC at that time) and it worked fine after that.

I tried running furmark overnight, it still runs it after about 7 hours, but well I got 75 degrees both cpu n gpu.

I tried prime95 for 10 minutes each mode, and 1 hour with most power-hungry mode, still none. I don't really have the time to check for everything at the same time so progress is a bit slow, sorry. I'm calling someone I trust to have a look on the pc, and maybe cannibal something from his spareparts. I just really hope it wasn't the gpu, mobo or cpu. also gonna try a really fresh windows install (100% full format and reinstall).

Oh, some things I have noticed :
-on win 10, csgo would crash twice every 5 minutes. but on win 8 (haven't got the win 8 anymore) csgo would only crash two times and then it'll run as smooth as a fully working rig, but with rare occasions of texture rip which goes back to normal that instant.
-bf1 still keeps crashing
-the longer I put the pc on power off mode the longer I could play games before it actually crash, but not on a notable length (fresh run after sleep : bf1 reached 20 minutes before purple/red/white checkers appear everywhere and the usual dxgi_device_removed thing pops. run after some work done : in the very first gameplay cutscene [not video cutscene] it already crashed)

I have found a 9600gt in my garage, my trusty friend tested it and it ran games as it should, and we will be testing it on tuesday. I got a week school break since easter is coming soon, and I'm going to put the week to a good use.

There will be 2 scenarios :
1. still artifacts then crash
2. no crashes anymore.
should I get one of those, what component in my rig should I pay more attention to? for example : if I get the first scenario I should test more on the PSU and CPU, and maybe mobo (motherboard).
If the second scenario is present there will only be little variables : either the GPU is doing an infant-bricking or the motherboard didn't support pascal yet, or something else which is related to GPU or it's support.
But if I get the first one there would be a ton of variables, example : bad CPU, bad PSU, bad mobo, bad Windows copy (unlikely), bad RAM (extremely unlikely since I have ran memtest to completion), bad slot, bad power distribution, etc etc.

 

[PotatoPCGamer]

Disconnect and reconnect the Card from the mainboard slot

tried that, and it didn't work

 

[helpstar]

What about the 9600GT?