[Krgwow]

Hi guys, i have the following specs:

Core i5 3570K 3.4 Ghz @ OC to 4.6 Ghz + Noctua NH-D14
asROCK Z77 PRO 4
GTX 980 G1 Gaming @ 1550/8000 with Gamestable BIOS
2x4 GB G-Skill DDR3 1600 Mhz
CX Corsair 600W PSU
SSD/HDD

Well... the games are:
Shadow of Mordor: my PC just freeze after a while, like 15 minutes after playing(i have the same problem with my old R9 290)

Far Cry 4: thats pretty random, one day i got the freeze after playing 7 hours and sometimes 5 min in game to get the my pc freeze, when i CTRL ALT DEL on the freeze, on my desktop i have the error: DEVICE REMOVED bla bla bla GTX 980 bla bla bla DXGI

Dragon Age Inquisition: pretty random too, exactly the same thing of Far Cry 4, freeze and device removed, etc...

Facts:

- I have a modified BIOS, the G1 Gigabyte Gamestable. The only game that i tested with the original Gigabyte BIOS was the Shadow of Mordor, and my game freeze exactly the same.

- I did a fresh Windows 8.1 install before install my GTX 980, did a fresh install of nVIDIA drivers twice(344.75 and 347.09) with DDU(Driver Sweeper/Driver Uninstaller, safe mode, clean registry, etc...)

- On every games my temperature are just fine, on GPU AND CPU.

- Other games like ARMA 3, Crysis 3, DayZ, apparently work well. But never played all that games for long time like Far Cry 4.

- Already tryed play with LESS overclock... 1530/7800, 1540/8000, and 1550/stock, nothing works. My Graphics card are with max voltage on MSI AFTERBURNER, prioritizing the temperature, no Link and temp target 91ºC(never get pass 71ºC playing). My CPU are OK too.

- Already start the games with EVERYTHING closed(especially MSI AFTERBURNER/Rivatunner), Windows updated, billions of DirectX versions i try to install, all the Microsoft C++ things, everything installed.

- Already stressed my Graphics Card on 1 hour or more on MSI AFTERBURNER GPU BURN test, max temp 83ºC(got some throttle, not much), saw some artifacts on Core 1550Mhz, on Core 1540Mhz not a single one). No CRASH, no Freeze, no black screen, nothing. The game that i use to stress my card is Crysis 3, didn't see any artifact playing on 1550/8000, only on Afterburner. But as i said, 1540 are ok, even on Afterburner.

- Already played Uningine Heaven Benchmark for like 3 hours or more no stop and nothing happens, everything went just fine.

- And the most curious, when i got that error on my card, ANY OTHER GAME THAT I PLAY i will get the same error BEFORE the game start, before MENU! it just gave me a black screen and the error shows up! i have to restart my PC to my games works again!

I'm running out of options here... is a GPU problem? should i RMA it? change my BIOS again?

 

[21cage12]

Lots of things to read but..An overclock can mostly cause certain games to freeze or crash, sometimes even if the overclock is stable, have experienced it myself. DX issues is prone to these problems as well, so i had make sure i am up to date from the bios to windows including my drivers; and re-installation of dx.

 

[ronnin426850]

You said you tried with LESS overclock, but did you try with NO overclock? Like, really NO oc - stock CPU, stock RAM, stock GPU? Just for troubleshooting purposes. Try it and report back

 

[deepor]

I'd look into the CPU overclock as you already tested things out with the graphics card, and you also say that you had the same crash with a different graphics card. You should test with the CPU running at BIOS default settings, see if the crashes still happen.

About the CPU overclock you currently use, can you find warnings with the source "WHEA-Logger" in the Event Viewer of Windows? That often helps with finding that things aren't actually completely stable even though prime95 or whatever runs fine.

 

[Krgwow]

just tested with no overclock at all on my i5 3570K(UEFI settings) and crashed like it was before, no sucess

already unplugged all my stuff too, celphone recharger, usb cable from my monitor, psp recharger, gamepad, no sucess, still crashes
only have connected my usb headphones, keyboard and mouse

well, i'm going to test with no overclock on my GPU then... damm

 

[Krgwow]

testing Far Cry 4 RIGHT NOW with:

Core clock: 1319 Mhz(1329 stock)
Memory: 3488Mhz(3500 stock)

brb

 

[Krgwow]

well, the game crashes two times on a mission when it was overclocked, now, a little below stock speeds the game works fine, did the mission no problems

that means i pushed the overclock to hard? or it just that games? thanks for the help btw guys

 

[KYKYLLIKA]

If you get stable results at stock clocks and crashes at overclock then you probably just pulled a losing ticket in silicon lottery.

HOWEVER… there is one distressing detail: the error reports that your card is removed, which should not happen with a simple overclock chip failure (it should be either throttling itself or generating corruptions leading up to a blue screen). This may as well indicate that your card stops getting power, at which point the suspect would be the power circuit on the card (since 600W PSU is plenty to drive the card, the CPU, the cooling, drives, and whatever else you can throw into a single GPU system).

If you have the facility to measure and record voltage/current (a good multimeter with continuity testing would be the tool), then you might want to check that out.

Another suspect would be not that the card stops getting power or fails, but that the driver falls off the cliff for some reason, forcing the system to report that there is no card, but still using the basic display facilities (kind of the same way you can use your card without drivers installed). That could happen for a number of reasons, like the card identification blowing (it has a small set of transistors or even a whole chip which serve as hardware identifier for compatibility purposes), that might indicate a circuit failure in a non-critical area of the board.

I wish these things were clear and accessible for us the users, but the user experience designers seem to have made it their goal to come up with the most incomprehensible and twisted error reporting and exception handling in their quest to create "user friendly" interfaces.

 

[Krgwow]

Quote:

Originally Posted by KYKYLLIKA

If you get stable results at stock clocks and crashes at overclock then you probably just pulled a losing ticket in silicon lottery.

HOWEVER… there is one distressing detail: the error reports that your card is removed, which should not happen with a simple overclock chip failure (it should be either throttling itself or generating corruptions leading up to a blue screen). This may as well indicate that your card stops getting power, at which point the suspect would be the power circuit on the card (since 600W PSU is plenty to drive the card, the CPU, the cooling, drives, and whatever else you can throw into a single GPU system).

If you have the facility to measure and record voltage/current (a good multimeter with continuity testing would be the tool), then you might want to check that out.

Another suspect would be not that the card stops getting power or fails, but that the driver falls off the cliff for some reason, forcing the system to report that there is no card, but still using the basic display facilities (kind of the same way you can use your card without drivers installed). That could happen for a number of reasons, like the card identification blowing (it has a small set of transistors or even a whole chip which serve as hardware identifier for compatibility purposes), that might indicate a circuit failure in a non-critical area of the board.

I wish these things were clear and accessible for us the users, but the user experience designers seem to have made it their goal to come up with the most incomprehensible and twisted error reporting and exception handling in their quest to create "user friendly" interfaces.

Could be the modified bios? i can't measure the graphics card voltage... i don't have a multimeter
I'm getting DirectX error on my Dragon Age by now and on Far Cry 4 doesn't show any error, just a freeze, have to end the task and restart my PC, otherwise my graphics card won't work

well... really don't know what to do... i'm thinking put my Core Clock on 1540 again and no Memory overclock, so i can see which one i push too much, if that's the case, of course

thinking about modify my bios again... it's ok put a "gamestable bios" from EVGA SC on my G1 Gigabyte? since the bios that i download on the topic disappear and that from EVGA took the place

ah! before i forget, on the first try to modify my GTX 980 my PC crashed and i have to start my PC with onboard graphics since my card has corrupted bios, i flashed the original again, then gamestable bios, and here i am
maybe i damaged the card? is that a possibility? tryed FURMARK TODAY with 1540/8000 exactly 47 minutes and no crash or artifact
if there is a problem with my PSU, Furmark should crash the PC, since stress my GPU more then any game on the market, right?

 

[Kidam101]

Psu not able to juice up ur setup when overclocked ??

Sent from my iPhone using Tapatalk

 

[Chargeit]

I'd look at that PSU.

You're pairing a budget PSU with mid - upper ranged gear with OC'ing.

 

[Farih]

Sorry to ask this qeustion but are you games legal ?
Hack/patched games can give alot of trouble sometimes and we are not allowed to discus/support them here on OCN.

If your games are legal then i would ask do you have C-states enabled in the BIOS ?
If so disable then.

It can be very tricky to get a 3570K @ 4,6ghz stable with C-states on.
Can explain why you have no problem in benchmarks because they load everything up to a 100%
Curtian games require less of a CPU and thus can make your CPU fluctuate in speed when you have C-states enabled.

 

[Krgwow]

since you guys think it's my PSU, then when i be at home, gonna test PRIME 95 and Furmark at the same time for 1 hour or so, then if the problem are my PSU, my PC will probably freeze at desktop, right?
i mean, that's make sense to me

 

[Chargeit]

I'd set everything to stock. Test the ram. If the ram does OK, then test your CPU/GPU.

I can say I had bad ram once that passed all ram tests. However, when I replaced it it fixed crashing issues.

I had a PSU that was bad (TX850M) that I was able to stress test fine. Though even after passing stress tests it was causing my games to crash and having random windows freezes. I blamed that on the GPU for a while. I finally replaced the PSU and haven't had a issue since.

I wish everything was cut and dry. But, at least in my experience sometimes it's hard to nail down what's the cause of your problem since sometimes the usual things don't expose it, or, can lead you to believe it's something else.

 

[Krgwow]

Quote:

Originally Posted by Chargeit

I'd set everything to stock. Test the ram. If the ram does OK, then test your CPU/GPU.

I can say I had bad ram once that passed all ram tests. However, when I replaced it it fixed crashing issues.

I had a PSU that was bad (TX850M) that I was able to stress test fine. Though even after passing stress tests it was causing my games to crash and having random windows freezes. I blamed that on the GPU for a while. I finally replaced the PSU and haven't had a issue since.

I wish everything was cut and dry. But, at least in my experience sometimes it's hard to nail down what's the cause of your problem since sometimes the usual things don't expose it, or, can lead you to believe it's something else.

i don't know... if the PSU was the problem i think the crashes will happen on every game that pushes my PC, like ArmA 3, Crysis 3... and neither of them happens anything

many people have the problem with DirectX on Dragon Age Inquisition, even the EA ask players to turn off OC on graphics card to play that game, since i refuse to turn of mine, i got the random crashes

on Far Cry 4... well, 5 patches in one month

anyway, gonna do more tests
is there any way to test a PSU without a multimeter? stress or something?

 

[bfedorov11]

A "kill a watt" outlet meter is great in this situation so you know your power usage. If I had to guess, its probably the psu.

I would flash the card back to the original bios. If you can't, try flashing it again or with a different bios.

If not, reset cmos, run stock. Tweak one at a time till the problems comes back.

There is no telling if its the psu unless you replace it or have a meter so you know how many watts you're pulling. Overclocked gpus pull a MASSIVE amount of juice.

 

[Krgwow]

Quote:

Originally Posted by bfedorov11

A "kill a watt" outlet meter is great in this situation so you know your power usage. If I had to guess, its probably the psu.

I would flash the card back to the original bios. If you can't, try flashing it again or with a different bios.

If not, reset cmos, run stock. Tweak one at a time till the problems comes back.

There is no telling if its the psu unless you replace it or have a meter so you know how many watts you're pulling. Overclocked gpus pull a MASSIVE amount of juice.

that's so weird, i play DayZ, ArmA 3, Crysis 3, FIFA 15, Metal Gear V, Insurgency, lot's of games with no problem
i had a defective PSU once on my 7800GS from EVGA and my PC simple restart when i was gaming anything, or give me a freeze that i have to hard reset my PC

maybe Far Cry 4 don't like OCed GPUs?

 

[Chargeit]

Yea, when my PSU was the issue it didn't always do it when you'd expect.

Get a cheap multimeter to test it out.

Personally, after dealing with BS like you've mentioned I'd just pull the parts I suspect of causing the issue. Setup RMA's, fight with customer service until they agree to pay for shipping, and send the parts in for replacements. Tell them they've already costed you enough time and money and you can't afford the shipping. (I'd look at RMA'ing the PSU and RAM)

If money isn't a huge issue then buy replacement parts before shipping your stuff out. Once you get the RMA'ed parts in you can sell them, or, keep them around as back up.

Sounds like a lot of trouble, but, it's better then spending the next 2 months trying figure out which is the problem part.

 

[Krgwow]

Quote:

Originally Posted by Chargeit

Yea, when my PSU was the issue it didn't always do it when you'd expect.

Get a cheap multimeter to test it out.

Personally, after dealing with BS like you've mentioned I'd just pull the parts I suspect of causing the issue. Setup RMA's, fight with customer service until they agree to pay for shipping, and send the parts in for replacements. Tell them they've already costed you enough time and money and you can't afford the shipping. (I'd look at RMA'ing the PSU and RAM)

If money isn't a huge issue then buy replacement parts before shipping your stuff out. Once you get the RMA'ed parts in you can sell them, or, keep them around as back up.

Sounds like a lot of trouble, but, it's better then spending the next 2 months trying figure out which is the problem part.

about the multimeter i will have to buy one, as for the memory, i have the exatcly same memory on my cousin PC's, 2x4 GB g-skill ripjaws ddr3 1600, will be easy to test it out
but since i play heavy games like Crysis 3(that i used only for OC tests) and ArmA 3(best game to test CPU OC

) i really dont think the problem is my PC, but i will do some tests when i get home

is there a problem stress a PC with furmark and prime95 at the same time?

 

[Chargeit]

Test the PSU with a multimeter. If that works out, see if your cousin will switch out ram with you for a week.

As long as you can cool and power it, then you should be able to run them. I wouldn't leave them unattended though. The idea of running that combo with a possibly defective PSU does worry me though.