Anyone got a handmedown graphics card they want to sell me!!

[Twisted]

Ok I've been having some serious issues with my PC of late BSOD/freezing during games. I've tested pretty much everything and was about to buy a new card but as I don't really know if its the gfx card or something else so I figured I'd see any of you like me keep hold of old hardware that you want to offload rather than me just blowing £300 on a potential paperweight and the only spares I have are my two old HD4850X2's so they don't really give me a good comparison.

My current card is HD7850 2GB DDR5 so i'm really looking for something similar to check the power draw etc. (I replaced the PSU a with a 1000 watt bastard from hell while back and even hooked the gfx card up to its own 600watt PSU and it still bombed out) and ideally an Nvidia card (I think the 660s are about the Nvidia equivalent the same range) as im seeing a lot of people saying they are getting the same issues with ATI cards. I also don't think its overheating as I can run certain things like benchmarking tools for days and there are no problems but wack in DayZ, Loadout or even Borderlands 2 and Its a game of russian roulette until it bombs out (spent 4 hours in load out yesterday and 15 minutes today)

Either bob me a PM with what you have/want for it or if you have some brainwave and know what the issue is let me know

[larchy]

All you're doing with a PSU with such a high wattage is lowering the efficiency as the power draw will be so low. What brand/make is it? (please don't say OCZ or thermaltake) The voltage regulation/current supply is the issue, not the wattage, and with such a high wattage running at what is almost inevitably such low efficiency the voltages are far more likely to go out of spec.

If you suspect a faulty card I don't see why the 4850x2 won't help you troubleshoot. 7850 TDP 130W, 4850X2 TDP 220W; 90 W isn't going to make any difference to power issues when you're going to have about 700W spare capacity anyway.

whats the RAM?

[Twisted]

All you're doing with a PSU with such a high wattage is lowering the efficiency as the power draw will be so low. What brand/make is it? (please don't say OCZ or thermaltake) The voltage regulation/current supply is the issue, not the wattage, and with such a high wattage running at what is almost inevitably such low efficiency the voltages are far more likely to go out of spec.

If you suspect a faulty card I don't see why the 4850x2 won't help you troubleshoot. 7850 TDP 130W, 4850X2 TDP 220W; 90 W isn't going to make any difference to power issues when you're going to have about 700W spare capacity anyway.

Whats the RAM?

Its a Corsair PSU and its actually only 860 the previous was 600 (I remember thinking 1000w was overkill at the time but was looking at both) The reason I wanted a similar card was to test as close as possible to the same settings etc.. if I'm honest I know for a fact its not the power as I hooked up the old PSU just to the gfx card once I put this one in and it still bombed.

Specs
Mobo - Asrock Z77 Extreme6
CPU - Intel Core i5 3570K 3.4GHz
RAM - 16GB Corsair Vengeance DDR3 1600Mhz CL9 1.5V Non-ECC Unbuffered 4x4 sticks

The setup I have is with 2 monitors one from the graphics card via DVI and the other is directly from the mobo using the inbuilt intelHD(I need new monitors really but can't justify that as I spend more time working that gaming these days and if I want 1080 gaming action I can just run the HMDI from the graphics card to the 50" TV and use that!) however when I did the reformat I didn't install the intelHD or the second monitor and ran the PC and again still died.

I was originally convinced it was the card being faulty I've tried some of the fixes people posted about the catalyst control center causing issues no luck, I even
reformatted and just in case it was corrupted drivers but after posting on another forum (sorry!!) they convinced me my old PSU was out of spec and old although I didn't think so (was made by Hiper) but again spending £50 on a PSU isn't the end of the world especially as I needed another one for my test environment anyway as that was running on some old 300w PSU that died in a blaze of glory due to overheating.
I also ran memtest on the RAM and there was nothing there, I couldn't find an app to test the GPU RAM.

I'll give the 4850 a go and try to stress it out and see if it BSOD's



The only external factor really is that the base unit is plugged into a UPS that shared with 2 Monitors and a NAS but again being logical if it was to do with that I would assume I would see the other devices fail more, plus the PC stays on and the sound keeps looping just freezes up, the CPU fan keeps spinning but the graphics card fans all stop and for all intents and purpose its dead until I reboot, its looks like an overheat but if i'm running a monitor it stays well within normal temps, even the GPU load doesn't seem to matter I've had it crash when its been running at 99% for half an hour and crash when its at 40% for 10 minutes.

[larchy]

It's not going to be anything to do with the power I'd say...

It's difficult trying to "stress" a GPU as benchmarks tend to stress them differently to actual games.

Just out of interest, what is the BSOD code? (admin tools>comp management>event viewer>windows logs>system and look for big red crosses)

These sort of temperamental issues can be some of the most difficult to troubleshoot :/

You know the first thing I would do though.... get rid of two of those sticks of RAM and give it a try.

[Twisted]

Nothing in event viewer though as it seems to die before it can write to it the BSOD pops up then dies and its just blank main screen and secondary screen sticking with what was last on it the only error i get is after the restart
"The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

which is about as useful as sex toy made from barbwire and why I'm drawing complete blanks on it as I can literally get no feedback from windows on the issue.

I'll give the RAM a go tomorrow and see if that gives me some stability.

[larchy]

Windows+break>advanced settings>advanced>startup and recovery>settings>system failure> uncheck "automatically restart"

Should give you a BSOD with a code rather than an instant restart.

[Brizie]

You know the first thing I would do though.... get rid of two of those sticks of RAM and give it a try.

[Twisted]

Holy mother of god Brizie! MY THREAD HAS THE POWER TO REINCARNATE THE DEAD!

Yeah I've messed about the the RAm modules and it 'seems' a little more stable but as its sporadic we shall see. Might leave Dayz running for an hour or 6 and see what happens.

[Cacker]

If you're going to find out what's causing the issue, I tend to rip out all unnecessary devices and leave bare bone items. So disconnect any USB devices, DVD Rom etc. I've had a DVD hang the PC, and it worked fine once I'd disconnected it. It must have been screwing up something on the IDE, locking up the interrupt line or something.

Only use one stick of RAM. Then start swapping out each part and test. It's not nice, but you could order a new GFX card online and try it out, return it if turns out NOT to be the GFX card. Anything bought online comes with a 14 day no quibble return policy. Same with Memory and PSU.

Before you do that, test your system parts.

Have you tried loading the CPU/Memory/GFX card separately independently ?

There are numerous free tools to do this. If you get the Prime95 app, this can load your CPU and or Memory.

There's another app called FurMark which can load just your GFX card.

You can soak test with these apps, and see which one triggers the problem. Only load one thing at a time.

I tend to use RealTemps and HWMonitorPro to check CPU and GFX temperatures.

[Twisted]

There are numerous free tools to do this. If you get the Prime95 app, this can load your CPU and or Memory.

There's another app called FurMark which can load just your GFX card.

Yeah I ran furmark for 4 hours not a single issue other than I got really obsessed with trying to find fury doughnuts (and no thats not even a euphemism!) I might give that Prime95 a go although Memtest didn't throw up any problems, running on two sticks atm so will see how that goes.

[larchy]

The thing with furmark, as I mentioned earlier, is "benchmarks" and suchlike like that don't always stress the GPU in the way a game does. GPUS are that complicated these days that furmark can quite happily "stress" a GPU but still leave certain parts of it unused; furmark isn't going to be using the tesselator for example, whereas BF4 most certainly will.

[Cacker]

That maybe true, but, I'm not sure if the gfx is fooked it's going to be a particular function that's failing. The furmark can get the gpu running at 100% load and near its max temps. if it's not utilising shader 3.0 or doing any particular effect, I don't think that's going to be a biggy is it ?

It's not a software bug we're looking for surely, it's a HW issue ?

[larchy]

I've seen GPUs fall over in particular games because a particular part of the GPU has been stressed that caused the voltage to go out of spec; furmark was absolutely fine on it. Was that a GPU, software or PSU issue? Well, a decent PSU fixed that.

It's not as simple as "100% load" - it never really is 100% if parts of the chip aren't getting used, and in reality you can never truly get to 100% with with the massive parallelism. Sure you can make the temp go through the roof by drawing eleventy-billion triangles, but when you look at how these things handle power draw, load balancing, power gating etc it's far more complicated than that, especially the most recent GCN Radeons that handle temp/power/clock balancing far differently than has been conventional in the past.

So yes, BF4 could well demonstrate a hardware failure point that furmark fails to. I'm not saying that's the case here, just pointing out the limitations of these tools when suggesting them.

Same with Prime95 that may well peg a cpu thread at "100%" but which is far from a true picture of exactly what it's doing on a modern CPU. Tbh there's been no reason to use it for years now there's a version of linpack out there.

[Twisted]

Well its definitely not the RAM.

Went to two sticks - crashed
Went to the other two sticks - Crashed
Went to one stick x4 - 4x Crashes

Even tried a stick from my other PC and still bombed out.


I'm now a professional at making it bomb out.

VLC on second monitor playing some random crap (Spaced usually)
Playing Chronicles of Riddick Butcher bay on main and without fail after 30 minutes or so boom and it's not like its a recent game either so it shouldn't be that taxing! I guess im just going to get a new Graphics card..not sure if I want to risk another Radeon but if it is the mobo and I need to replace that at least I can Crossfire them if they both work....choices choices.

[Twisted]

Well new card and still getting issues, I think I might be having an unlucky year. Got a new card (Toxic R9 270X) all going fine then boom instead of it freezing it now gives me serious artifacting after a while even when I quit and go back to the windows desktop its only running a temp of 38 degrees, im hoping its just the new card is fecked and I got unlucky time to RMa it I think.

[Kiwi]

It's definately the.... Who am I kidding I don't have a clue. Hope you get it sorted it's frustrating when you can't work out the problem.

[Fluke?]

Loving the Win+Break combo! I don't think I've ever used the Break key on my keyboard plus my love for keyboard shortcuts has made my day
Plus I get a similar issue why my machine will suddenly reboot without a bsod so this should be doubly helpful!

[larchy]

So it's failing, but in a different way. You have no way of determining if this is related to an underlying fault with another component, or if you have two faulty cards. Without a second PC to test with it's just a crapshoot of best guesses and I don't think you're going to get anywhere.

If a third GPU doesn't resolve the issues then it would suggest something else is at fault, but tbh what do you do next? Is it certain that the software is not at fault? Did you get anywhere with BSOD codes? Change the motherboard as the likeliest culprit? What if that's not the problem? How confident are you of your installation? No soft-grounds anywhere? You need a known good set of components to test with and a clear methodology else you're going to go around in circles and end up with a lot of unnecessary expense and wasted time.

[Twisted]

Plus I get a similar issue why my machine will suddenly reboot without a bsod so this should be doubly helpful!

Overheating somewhere would be my guess.

[Fluke?]

Overheating is a sensible theory but when mine goes it is usually a few minutes after the first boot and usually when I'm just starting to click around a few Brazilian fart porn websites. After that it will often give a Buuuuuur Bip at post and the behave normally even when playing games for any lenght of time.
Writing this down makes me question why I've never actually spent any time trying to sort it out but I guess I'm just lazy and have gotten used to it.