5870 gray screen crash.

I fixed it with removing my PCI-e soundcard, so now videocard PCI-e slot can run at full 16x speed, instead of 8. Haven't gotten 1 crash with a grey screen (&weird sound) since then! Normally I would get it with certain games in less then 20 minutes. Using a p55 motherboard btw. Luckely my onboard sound isn't that bad, so I can live with this..


[edit: just had a crash again, but that took about 10 hours playing]
 
Last edited:
Don't have the greyscreen problem, but the 2D/idle/desktop crashes are no longer coming up for me on my 5770 after switching the Windows desktop theme from Aero to Basic. It's been a week now, and not one crash. Touch wood.
 
does someone know if ati will be fixing an issue with the 2d flcker with dual monitor setup and enabling overdrive, when I disable overdrive the issue is gone because the clocks go back up to 400 as required for dual monitor setup, but when I enable overdrive the clocks go down to 157 on the core and thats when the flickering starts.
 
I replaced an XFX 5770 with a Diamond 5770 and it's been several days now with no crash. I didn't reinstall drivers or anything, just swapped the card and let it auto install the same driver, then rebooted. Could this be related more to brand of card? I know XFX is very popular with the warranty.
 
Powercolour 5850 here and no probs, does any 1 have any probs with a powercolour card?
 
NKD said:
does someone know if ati will be fixing an issue with the 2d flcker with dual monitor setup and enabling overdrive, when I disable overdrive the issue is gone because the clocks go back up to 400 as required for dual monitor setup, but when I enable overdrive the clocks go down to 157 on the core and thats when the flickering starts.
Click to expand...

You can do all the same with MSI Afterburner, and you card won't downclock either...

It works for me anyway, and I have dual monitors...
 
geekusa said:
x2! That's my same exact problem to the letter. With 10.2 I would get the GSOD, with 10.3 I was getting the BlackSOD.

Also 120MHZ refresh rate is very flaky on the 5870 & 5770.
Click to expand...

Make that x3. I have the XXX version of XFX 5870. I'm trying to get them to RMA now. My clocks (after BIOS flash) is 900/1250, where before was 875/1300, where it would hard lock completely.
 
I have gone through 2 cards, I dont Overclock and i still get grey screen lockups or artifacting and end up having to reboot the pc.... im getting sick of this, ready to take my card back to frys for a 3rd time and try to get an nvidia or something....
 
also, its not the pc I dont believe because when i get a grey/red screen I can still hear audio looping in the background... Im using onboard sound. and I have an 850 watt pc&c psu
 
So here it is, a month later, and I'm still having the very same issues with this ATI 5870 as I was having three to four months ago. Kyle contacted AMD, AMD contacted me, AMD contacted ATI, ATI contacted me, and they just tried more troubleshooting. No replacement card, nothing. ...and here I am, still stuck with this POS 5870 from a wretched company called XFX. I will NEVER buy XFX again, and I recommend everyone else stay away from them as well. $440 paperweight... right here. I wish I knew a good lawyer that dealt with this type of stuff. It crashes in this PC constantly, and won't it work in my other gaming system at all. I tried it in my mom's computer and it crashes just like it does in mine.
 
did you contact XFX? when i did they told me to try a couple of troubleshooting solutions and said if those didnt work i could RMA it
 
Corban227 said:
did you contact XFX? when i did they told me to try a couple of troubleshooting solutions and said if those didnt work i could RMA it
Click to expand...

Since he's been in contact with AMD, you'd think AMD would get XFX to RMA, or AMD would just RMA it themselves?

I still have black screen of death issues with my 5870 about every 48 hours now. GSODs were gone with 10.2, but black SODs are now more common than ever with 10.3...

i think the 10.3s made my black screens worse... I could go almost a week, now I can't go 48 hours...

Since the drivers have all had some effect pro or con on my death screens, mine still appears to be a driver problem...

If I finally come to conclusion I'd best just RMA it, and can't get XFX or AMD to do so, my last resort will just be to never buy an AMD product again...
 
Hey Tolyngee, I'm just curious, what motherboard do you have?

Tolyngee said:
Since he's been in contact with AMD, you'd think AMD would get XFX to RMA, or AMD would just RMA it themselves?

I still have black screen of death issues with my 5870 about every 48 hours now. GSODs were gone with 10.2, but black SODs are now more common than ever with 10.3...

i think the 10.3s made my black screens worse... I could go almost a week, now I can't go 48 hours...

Since the drivers have all had some effect pro or con on my death screens, mine still appears to be a driver problem...

If I finally come to conclusion I'd best just RMA it, and can't get XFX or AMD to do so, my last resort will just be to never buy an AMD product again...
Click to expand...
 
Kelthuron said:
Hey Tolyngee, I'm just curious, what motherboard do you have?
Click to expand...

Gigabyte GA-EP43-DS3L

Know something I don't?

Not being some damn fanboi, but the system was rock-solid stable OCed with an Nvidia card in it. I still get black screens with the 5870 with my system un-OCed, and apparently even more often with the 10.3s than I did with the 10,2s,,,

No GSODs since 10.2 though...
 
I'm using a msi 4870, (currently) 9.10 drivers, Win7 x64 running DX11 on a 2.83 quad lga775 board. I've been getting black SODs for about a week now since I reinstalled the OS on my SSD. Before then, I had absolutely no problems with graphics, not I can play anywhere from 20 minutes to 3 hours before it crashes without warning. In a couple games, the screen will freeze, go black, and restart itself and I keep playing. In HL2 it did this 3 times over the course of 40 minutes of play before finally BLSOD and I had to hard-reboot. The GPU temp stays at about 63, maybe as high as 65 during gameplay.

I've tried using 10.2 and 10.3 drivers, OC, under-clocking, nothing has worked so far. But this isn't a 5xxx series issue only! I've seen many posts of 4xxx users saying they've gotten the same symptoms. If anyone has advice, I'd appreciate it though.

Another poster said
I watercooled my 5970 which would constantly GSOD within 2 to 5 mins of playing any 3d game eyefiinty or not. No issues with desktop or 2d use. I thought it could be a cold bug.
Click to expand...
I've found this to be about similar in my experience: HL2, ETW, HoI3, Cellblock Squadron, M&B all crashed at some point w/ the black screen (which window's error says is a blue screen when I look at the details). But I ran Gratuitous Space Battles (intensive 2d) for 3 hours, no problems before I quit the game to go to sleep.
 
Kelthuron said:
So here it is, a month later, and I'm still having the very same issues with this ATI 5870 as I was having three to four months ago. Kyle contacted AMD, AMD contacted me, AMD contacted ATI, ATI contacted me, and they just tried more troubleshooting. No replacement card, nothing. ...and here I am, still stuck with this POS 5870 from a wretched company called XFX. I will NEVER buy XFX again, and I recommend everyone else stay away from them as well. $440 paperweight... right here. I wish I knew a good lawyer that dealt with this type of stuff. It crashes in this PC constantly, and won't it work in my other gaming system at all. I tried it in my mom's computer and it crashes just like it does in mine.
Click to expand...

RMA dude. My first XFX 5870 was junk but my replacement works fine. Just bad luck.
 
Just curious if the RMA had helped? I just bought a 5870 Vapor-X, and I can't seem to play a game longer than 5 min without either a grey screen, black screen, or screen with veritical lines freezing up my system.

It's pretty frustrating, and I'm unsure what to do at this point. I'm using 10.3 drivers, and this is a fresh build, so it should be pretty clean.

Running the following:
i7-860
gigabyte p55a-ud4p
4GB Gskill ripjaw 7-7-7-24
X-fi Fatality Pro
D-link DWA-556 Xtreme N wifi adapter
Dell 3007WFP (gaming at 2560x1600)
win7 64-bit

I think that's all of the pertinent info.

I've had the 5870 for about 5 days now, so I can still return it if necessary, but I'd prefer not to eat the 15% restocking fee (bought from NewEgg) which would end up costing me about $65 or so.

If there were a fix, that'd be perfect, but it seems that there isn't. I've heard some people just RMA the card in hopes of getting one that doesn't have issues? If this is the case, I think I just have to pay for return shipping which should be less than $10. Just kinda sucks that this happens, and never figured I'd have all of these grey screen issues that I read about since so many other people/reviews keep praising these cards.

Any thoughts/ideas?

Thanks!
 
Kelter said:
If there were a fix, that'd be perfect, but it seems that there isn't. I've heard some people just RMA the card in hopes of getting one that doesn't have issues? If this is the case, I think I just have to pay for return shipping which should be less than $10. Just kinda sucks that this happens, and never figured I'd have all of these grey screen issues that I read about since so many other people/reviews keep praising these cards.

Any thoughts/ideas?

Thanks!
Click to expand...

Depends on why you have issues. I had grey screen error and the best fix was to use two screens instead of one. Not long ago, my PSU crapped out and took the card with it. I had tested this card with two power supplies in two PC's and the grey screen error occurred in both. Therefore I had ruled out PSU in troubleshooting earlier. I got myself a new card and I got grey screen error again. My PSU was also getting noisy and it turns out that both PSU's were failing. So, I bought myself a Corsair TX750W to hold myself over while RMA'ing the PSU's. No grey screens whatsoever.

My cards were probably never the problem and the grey screens were probably symptoms of failing PSU's.
 
you guys that are running i7 CPUs (dual or triple channel RAM), try running your RAM at 1.6 or even 1.65V and see if it stops the grey screen crash. I am just curious.
 
I suppose it could be a PSU issue.

I'm using a new XFX 650W XXX power supply that was on the HotDeals forum about 2 weeks ago. I was under the impression that 650W should be plenty for a single card 5870. I could use the PS from one of my other systems, but it is only a 550W Antec Neo Power.

Maybe I plugged in the power to the card wrong? The PSU comes with 2x 6pin PCIe connectors and 2x 6/8pin PCIe connectors built in. It also has 2 additional slots for modular 6/8 pin PCIe connectors. I just used the 2 built in 6pin to reduce cabling. Does it matter which ones I plug these into?

As to the higher voltage on the memory, I suppose I could try that as well, or maybe set the timings down a bit.
 
I read info from Corsair and even an Intel BIOS engineer recommend running i7 RAM at 1.6 to 1.65V for maximum stability. I know this fixed issues with my rig, but wasnt able to test my 5870 because I was so frustrated I went back to nVIDIA.

Its worth a shot! :)
 
Thanks for the advice. I just tried the voltage at 1.6, and still no luck. I even reduced the timings from 7-7-7-24 to 9-9-9-24, and BC2 and MW2 both crash within 5 min still.

I'm thinking it's just the card. I'll just RMA for a replacement for now I guess. Once that fails, I'll just return the card and go back to nvidia. Never should have left the tried and true :(

Trying to save a few bucks going ATi is now ending up costing much more than going nVidia once you factor in time wasted.
 
Kelter said:
I suppose it could be a PSU issue.

I'm using a new XFX 650W XXX power supply that was on the HotDeals forum about 2 weeks ago. I was under the impression that 650W should be plenty for a single card 5870. I could use the PS from one of my other systems, but it is only a 550W Antec Neo Power.

Maybe I plugged in the power to the card wrong? The PSU comes with 2x 6pin PCIe connectors and 2x 6/8pin PCIe connectors built in. It also has 2 additional slots for modular 6/8 pin PCIe connectors. I just used the 2 built in 6pin to reduce cabling. Does it matter which ones I plug these into?

As to the higher voltage on the memory, I suppose I could try that as well, or maybe set the timings down a bit.
Click to expand...

Your 550W Antec should be sufficient to test the PSU with. It wouldn't make a difference if you use the modular cables instead of the built in ones. Even though the PSU is new, it might not be 100% and your Antec is tried and tested.

I would also first try to turn off HDMI sound (disable it) to rule out those beautiful X-FI conflicts.
 
When I bought the MSI Hawk 5770 from the FS/FT thread here, I had black screen crashes where the computer would freeze or stall. It usually did this when a video was playing (MKV, MP4, AVI, or Youtube).

Various sources pointed to Flash 10.1 as the culprit, another pointed at RAM voltage, another pointed at HDMI audio on the 5xxx cards, but following either recommendations didn't fix the problem.

Now, I've gotten this video card just two days ago, and is in great condition. (Thank you, Cook_1) At first, I thought it was my power supply-- Coolermaster RS-500 500W PSU. But, it's been stable for over two years so it can't be it. It's powered 3 motherboards, three different processors-- AthlonXP 2800+ and Athlon X2 5000+, and now Phenom II X4 805. It's also powered as many as 4 hard drives and 2 optical drives, plus two different video cards-- 8600 GTS and HD3870. I haven't seen it fail yet. *crosses fingers*

I ended up bumping up RAM voltage from 1.8 to 1.85, but it still crashed during movie playback.

So, I did some searching around on Google and came upon this:
http://beecherbowers.com/2010/03/10/fixed-ati-5770-driver-stopped-responding-error-in-windows-7/

He gave a good explanation of what's probably causing the issues:
"Every time I have encountered this error, it has been when my system has been sitting idle or had the monitor powered down, then tried to resume it.

Windows 7 can call on your graphics card to manipulate the Aero interface. For some reason, the drivers take too long to ramp up the “idle” clock on the card to meet Windows 7’s demand. Windows 7 recognizes this as a timeout and restarts the driver."​

I had his same issue where "atikmdag" would just stop responding and I'd get a dozen or so of them logged in Event Viewer until I'm forced to do a hard reset on my computer.

What I eventually did, instead of following his instructions on making a profile so I don't have to load it every time, was to edit the BIOS directly.

For this, I needed two things: Radeon BIOS Editor (RBE) and ATI Winflash from TechpowerUp:
http://www.techpowerup.com/downloads/1770/TechPowerUp_Radeon_Bios_Editor_v1.25.html

http://www.techpowerup.com/downloads/1750/ATI_Winflash_2.0.1.11.html

I extracted both into the same folder and ran RBE.

Next, I selected "Acquire/Flash" and went to Browse to find the Winflash EXE (ATIWinflash.exe).

After that, I hit analyze, selected my card (which I only have one 5770) and hit Acquire BIOS.

Then, go to Clock Settings and look for "Clock Info 04." You can tell it's the right box when you see the lowest card settings during idle mode. Here, I set GPU to 250 MHz and RAM to 500 MHz, and kept voltage at 0.95.

I then went back to "Acquire/Flash" and instead of Acquire, I hit Flash BIOS.

Restarted, and it has been stable after playing five videos on my computer without atikmdag crashing on me. No more black screens.

It might work for other Radeon 5xxx cards by changing the IDLE state frequencies and (possibly) voltage for the GPU.

(Credit to "bbowers" for the suggested fix.)

Last note: Black screens were present in driver 10.2 then 10.3, and I eventually went to 10.3b Preview and it still crashed. So, it doesn't look like it's a driver issue at all, but a hardware issue.
It is now stable with IDLE clocks at 250 Core, 500 Mem, 0.95 V, and 10.3b Preview drivers.
 
Last edited:
Ugh.

System recently lasted two weeks without problems (only two weeks as I had to turn it off for other reasons...), and hadn't ever had a problem gaming...

But started playing TDU over the weekend...

This morning, card starts out at 875 (clocks), goes down to 600, then down to 400... 100% GPU usage being reported, so no reason to underclock itself... And framerate was definitely suffering due to this...

But then... BAM! GSOD!

Considering the fact that I'm four hours (not straight hours though...) into TDU (the game keeps track), I'd like to know WTF's up! I was at most 10 minutes into my current gaming session...

Yesterday I played almost three hours straight... No problems...
 
Tolyngee said:
Ugh.

System recently lasted two weeks without problems (only two weeks as I had to turn it off for other reasons...), and hadn't ever had a problem gaming...

But started playing TDU over the weekend...

This morning, card starts out at 875 (clocks), goes down to 600, then down to 400... 100% GPU usage being reported, so no reason to underclock itself... And framerate was definitely suffering due to this...

But then... BAM! GSOD!

Considering the fact that I'm four hours (not straight hours though...) into TDU (the game keeps track), I'd like to know WTF's up! I was at most 10 minutes into my current gaming session...

Yesterday I played almost three hours straight... No problems...
Click to expand...

Load up GPU-Z, play games in a window, and tell us what your VRM temps are. 1st reason I can think of why the card would start underclocking itself is overheating VRMs.
 
Spare-Flair said:
Load up GPU-Z, play games in a window, and tell us what your VRM temps are. 1st reason I can think of why the card would start underclocking itself is overheating VRMs.
Click to expand...

You'd think the fan would go above ~60% then...

I did find it odd that the fan would jump up in speed when I alt-tabbed out of then game to the desktop to look at a jpeg of a TDU map...

So: At the desktop, fan will go to 100% (or wherever it needs to go), but inside the game the fan will slow back down to ~60%.

I was starting to think I was hearing things, but MSI Afterburner showed that what I thought I was hearing was exactly what I was hearing...

FWIW: Doom3 seems to cause my card to work backwards: my temps are higher at the desktop (with the fan running harder, too...) than inside the game, where my temps go down, and the fan slows down...


The only oddity today was the GSOD... the other oddities are the norm... My card runs cooler when it's gaming... (and again, it's not due to the fan, as the fan slows down inside a game than when at the desktop...)

But to clarify... My fan will run at 100% if I alt-tab out of a game to the desktop... It's the only way it seems to be able to do it, though... So, if it's overheating, it's only when it goes to the desktop that it realizes it...

And again: yesterday it worked this way for three straight hours without issue, and another hour overnight... Early this morning it couldn't last 10 min tho...


Maybe I should just RMA my card... It's always acted like powerplay's working backwards... Running cooler when gaming than when at the desktop doing nothing...
 
I'm starting to think this is more and more of a heat issue on the 5870.

Originally, my wifi card was directly under my 5870. I moved it down to the second PCIe 16x slot instead to allow for more airflow to the 5870 and potentially less interference with the 5870.

After doing this, I was able to play BFBC2 for 2 full rounds with no crash (about 30 min or so). Needless to say, I was excited. Then, I switched to MW2 and boom. Crash within 1 min of playing.

From what people are saying here it seems that the fan speed on the card isn't working all too well if at all.

I will still try the PSU change when I get home tonight, but I suppose easier will be to try installing CCC and turning the fan to run at 80% all the time. Kind of defeats the purpose of having a quiet card like the vapor-x, and if this is infact the solution, I think I will just return it to get the Fermi.
 
check to see if your card has hynix memory on it
if it does get ahold of the manufacturer and tell them u need the newest bios
they will prolly ask for your part number and serial number

so if you have hynix memory this will fix your card because i went thru the same thing
 
Tolyngee said:
You'd think the fan would go above ~60% then...

I did find it odd that the fan would jump up in speed when I alt-tabbed out of then game to the desktop to look at a jpeg of a TDU map...

So: At the desktop, fan will go to 100% (or wherever it needs to go), but inside the game the fan will slow back down to ~60%.

I was starting to think I was hearing things, but MSI Afterburner showed that what I thought I was hearing was exactly what I was hearing...

FWIW: Doom3 seems to cause my card to work backwards: my temps are higher at the desktop (with the fan running harder, too...) than inside the game, where my temps go down, and the fan slows down...


The only oddity today was the GSOD... the other oddities are the norm... My card runs cooler when it's gaming... (and again, it's not due to the fan, as the fan slows down inside a game than when at the desktop...)

But to clarify... My fan will run at 100% if I alt-tab out of a game to the desktop... It's the only way it seems to be able to do it, though... So, if it's overheating, it's only when it goes to the desktop that it realizes it...

And again: yesterday it worked this way for three straight hours without issue, and another hour overnight... Early this morning it couldn't last 10 min tho...


Maybe I should just RMA my card... It's always acted like powerplay's working backwards... Running cooler when gaming than when at the desktop doing nothing...
Click to expand...

Fan speed operation has nothing to do with VRM temperatures, they are not connected. Fan speed is connected to the core temp, not the VRM temps directly. VRM temps change very rapidly and there isn't enough time for their temperature changes to really register in the primary heatsink. On many non-reference cards, the VRMs are even a separate heatsink altogether.

If you think the fan is the issue, just setup a hotkey profile for fan on 20% and fan on 100% so you can just turn on the fan with a few keystrokes. For example, I turn my fan to 100% by hitting, CTRL-ALT-SHIFT-Z
 
Spare-Flair said:
Fan speed operation has nothing to do with VRM temperatures, they are not connected. Fan speed is connected to the core temp, not the VRM temps directly. VRM temps change very rapidly and there isn't enough time for their temperature changes to really register in the primary heatsink. On many non-reference cards, the VRMs are even a separate heatsink altogether.

If you think the fan is the issue, just setup a hotkey profile for fan on 20% and fan on 100% so you can just turn on the fan with a few keystrokes. For example, I turn my fan to 100% by hitting, CTRL-ALT-SHIFT-Z
Click to expand...

Yeah, I went ahead and went back to just manually putting my fan speed to 100% before I game. Hadn't been an issue in a while.

Before that, I tried again amd watched what everything was doing, and the fan profile I put in MSI Afterburner wasn't working right (it should have been going up to 90-100%, but the card in the meantime would decide to downclock, and it appears everything was just confusing each other), but neither does just leaving the thing on Auto.

After what I saw there, my assumption is the GSOD may have been an overheat lock-up. Not sure why the previous day it worked hours without issue.

While it works, I shouldn't have to remember to manually adjust my fan speed. But that seems to work. IMO powerplay and everything else at Auto just doesn't seem to work right on these ATi cards... I witnessed that instead of increasing fan speed, it would just decide to downclock the card. "Auto" should have no problem with letting the fan speed go above 70% before it starts freaking out and downclocking.

At 100%, my GPU ran at about 57 celcius.


If that VRM temp was the issue, then again something's just wrong with the design?

Either way, the card doesn't seem to automatically control its own temps...
 
Yep, I just downloaded CCC, and set the fan speed to 80%. I was able to play BFBC2 for about 45 min with no issues. I switched over to MW2, and was able to play for about 10 min before I got the vertical screens. However, this particular vertical screen was different. I got the lines, and then after about 15 sec, the game flashed back on again. Then went back to the lines.. then 10 sec later, the game came back on, then lines, then it sounded as if my card just reset itself and turned off completely and back on, and then the game came back on and I was disconnected from the server and in the menu system of MW2. Everything was fully responsive at that point.

At any rate, what I saw from my brief testing was that the GPU temp was around 59C. I'm fairly convinced now that it is just a heat issue with my card. I will be returning the card today and looking for a GTX 480 replacement. If my card is going to be loud to be useful, I may as well get something else instead of spending the extra money on a Vapor-X edition.
 
So just an update. I got my gtx480 in today, and so far so good. So it looks like all the other equipment in my system was ok.

I was able to play BC2 for about 30 min with no issues.. occasionally on the 5870, I would get intermittant vertical line screens before the crash.

MW2 also seemed solid. Played for about 20 min with no issues.

Of course nothing has been extensively tested since I just got the card, but things are looking much better.

I'm hoping I just got a defective 5870 since I see many people that seem to have no issues with those cards, but from the number that I have seen that do have similar issues, something just seems very wrong.

Regarding power draw, I only run a single 480, no SLI, and no plans to ever SLI, and my 650W XFX XXX edition handles it fine.

The one issue with the 480 which has been mentioned quite often here is the sound it generates. It is significantly louder than the 5870. Even idle, it is significantly louder. Just browsing the internet, or posting this message, I can hear the 480. Is it unbearable.. no.. but it is noticable... similar to my old 9800GTX.

When running games, the 480 gets even louder. The 5870 was still VERY quiet. However, MY 5870 wasn't stable and I had to up the fan speed to 80% to be somewhat reliable in BC2, and let me play a round in MW2 before crashing. At 80% fan speed, it was louder than the 480 got.

Anyways, those are just my experiences. Sound is a big factor for me, and I am still a little undecided which card is better for me (assuming I can get a 5870 that works with my system). Most likely I will just keep the 480 since I have it already and it works, but I'm hoping that they come out with something similar to the Corsair H50 for the 480. I think CoolIT has something in the works but not available just yet.
 
I had some weird issues with TDU.

Even putting the fan at 100%, the fan would slow, and then the card would start to overheat. Alt-tabbing out of game would get the fan back to 100%, temps would drop, go back to game, things are fine... Then fan drops in speed again, temps go up...

A reboot fixed this behavior... but why!?

Now putting the fan at 100%, it stays at 100%.

But I was getting (first for me) constant GSODs, or BSODs. Sometimes within five min of loading the game, sometimes just when trying to load the game. Heat wasn't the issue.

If you can get past the 5-10 min mark, the game's then rock-solid, along with the desktop.

Hard to shout defective hardware and RMA when reboots eventually fix the issue.

While the GSOD issue sucks, the fan problem makes no sense at all.

I've played 29 hours of TDU. If you can get past the five-min mark, everything's rock-solid!
 
Haven't posted in this thread for a while but am doing so as I've experienced some GSODs in the last few days after months of no issues.

First off I'm using one of the leaked 10.5s or whatever they're supposed to be. So I can't rule out that being the cause. I'm starting to think it's got something to do with the core being at full usage. I've experienced it most with mount & blade: warband which in eyefinity and setting a max of 300fps (gets >100fps in game) saw the GPU under full usage all of the time. Under those settings I was getting grey screens pop up for a second a few times which is a warning that the crashing GSOD is on it's way. I eventually got one and it was a constant GSOD for a couple minutes. Then the game screen game back on but I wasn't on the game server anymore. Quit the game and the windows desktop is going nuts with all kinda artifacts. So reboot was needed.

I limit warband's fps to 100 which sees the GPU usage between 90 and 100%. Just had a grey screen flash for a second whilst in the game so decided to quit before the big one comes. Checked my GPU usage straight away and when that grey screen flashed on for a second the usage was at 100%. I also had a brief grey screen flash on in BC2 earlier. Wouldn't surprise me if the GPU usage at the time was 100% as well. My temps are fine as they max out in the early 70s (custom fan curve with afterburner).

Anyway I can't rule out these leaked drivers as the cause. It could well be them as they're the only thing I've really changed in the last few months of GSOD play. Just my obversations.
 
I finally figured out my 5870 issues...

Long story short, my Core Voltage was set wrong... Don't know how... Somehow it was at 1349, when it needed to be at 1350...

Now, I didn't set it at 1349, but for some reason it was there... And now that I've used MSI Afterburner to set it to 1350, I can't seem to find a way to force it to 1349 again... If I use MSIA to set it to 1349, when I hit "Apply", it goes right back to 1350...

So I don't believe I caused this, even accidentally...

But with it at 1349, the core voltage when gaming would go to 1.265, not 1.350 or 1.349...

Core Voltage at 1.265 apparently has a very drastic effect:

GSODs
fan doesn't work properly
PowerPlay completely freaks out


In gaming, when the CV would go to 1.265, the fan couldn't stay at 100% more than a few minutes... The fan would slow (even on auto), and the temps would quickly rise... This behavior was so bad that both MSIA and CCC were reporting that sometimes my fan was at 0% (yes, that is ZERO PERCENT!)

The fan would completely stop.

Alt-tab to desktop, fan immediately jumps to 100%... Alt-tab back to game, within seconds the fan would slow... and slow... sometimes again, to a stop...

Since you can't have the fan completely stopping, I now knew something was terribly wrong... But since alt-tabbing out of the game would put the fan speed immediately at 100%, it obviously wasn't a hardware or fan issue...


Now TDU's butter-smooth at 70+ fps... (2560x1600)

Doom3 (which a few months ago was sometimes a slideshow, and I thought it was just PP going nuts...) tested it at what was previously the worst spots on my 5870... Butter-smooth 60 FPS (game is capped at 60... beats the ~13FPS I had been seeing...)


So, if your card it absolutely freaking out on you, check all of the settings for everything...

Like I said, I don't believe I even accidentally put the CV at 1349, and the fact I can't seem to force it back to 1349 seems to say I'm correct...

I'd really like to know how it got there, and how long it's been like that... I bought a ton of games during the Steam holiday sale, and those games ran fine... So, 9.12 didn't do this... But by 10.1 or 10.2 this setting was off...

(I'd also like to note I played Portal in March without ever an issue... Only game during that time to never experience an issue, though... No idea why Portal's special...)
 
Tolyngee said:
I finally figured out my 5870 issues...

Long story short, my Core Voltage was set wrong... Don't know how... Somehow it was at 1349, when it needed to be at 1350...

Now, I didn't set it at 1349, but for some reason it was there... And now that I've used MSI Afterburner to set it to 1350, I can't seem to find a way to force it to 1349 again... If I use MSIA to set it to 1349, when I hit "Apply", it goes right back to 1350...

So I don't believe I caused this, even accidentally...

But with it at 1349, the core voltage when gaming would go to 1.265, not 1.350 or 1.349...

Core Voltage at 1.265 apparently has a very drastic effect:

GSODs
fan doesn't work properly
PowerPlay completely freaks out


In gaming, when the CV would go to 1.265, the fan couldn't stay at 100% more than a few minutes... The fan would slow (even on auto), and the temps would quickly rise... This behavior was so bad that both MSIA and CCC were reporting that sometimes my fan was at 0% (yes, that is ZERO PERCENT!)

The fan would completely stop.

Alt-tab to desktop, fan immediately jumps to 100%... Alt-tab back to game, within seconds the fan would slow... and slow... sometimes again, to a stop...

Since you can't have the fan completely stopping, I now knew something was terribly wrong... But since alt-tabbing out of the game would put the fan speed immediately at 100%, it obviously wasn't a hardware or fan issue...


Now TDU's butter-smooth at 70+ fps... (2560x1600)

Doom3 (which a few months ago was sometimes a slideshow, and I thought it was just PP going nuts...) tested it at what was previously the worst spots on my 5870... Butter-smooth 60 FPS (game is capped at 60... beats the ~13FPS I had been seeing...)


So, if your card it absolutely freaking out on you, check all of the settings for everything...

Like I said, I don't believe I even accidentally put the CV at 1349, and the fact I can't seem to force it back to 1349 seems to say I'm correct...

I'd really like to know how it got there, and how long it's been like that... I bought a ton of games during the Steam holiday sale, and those games ran fine... So, 9.12 didn't do this... But by 10.1 or 10.2 this setting was off...

(I'd also like to note I played Portal in March without ever an issue... Only game during that time to never experience an issue, though... No idea why Portal's special...)
Click to expand...

Whoa wait what? 1350mV? Mine is set to 1149mV and I believe stock is 1175mv?
 
You must log in or register to reply here.
Back
Top