A few months ago, against my better judgement, I purchased a P170SM with an AMD Radeon HD8970M installed. After a few days of operation my GPU driver crashed. At the time I thought it was the hot weather, but it happened again a few days later, so I started looking into possible software solutions like miscellaneous user settings, alternate driver versions, BIOS, VBIOS, underclocking, over- and undervolting, and changing my operating system from Win7 to Win8. It has persisted with every configuration I've tried.
I have not yet tried the open source drivers for Linux. It might be worth trying, but only as an experiment.
The symptoms are as follows: At random intervals, ranging from minutes to days, the screen freezes for 10-40 seconds. It then recovers with a standard popup message(Display Driver Stopped Responding and Has Recovered). Unfortunately, graphics context is never recovered, causing a crash in most programs(a notable exception is StarCraft 2, which has a recovery mechanism for this).
This only happens when the discrete GPU is in use, but it can sometimes be in use for many hours without incident. It is independent of temperature; I've seen it happen when the sensor reads as low as 45 celsius and above 70 celsius(in GPU-Z). It happens with graphically intensive as well as non-intensive software. It does not occur more frequently during stress tests, leading me to believe that this card would pass a factory test, but it is unusable for my purposes.
Now, I'm aware many of you will think that my way of going about things isn't very smart from a personal-economical perspective, but that is not why I'm here.
I would like to know is if this is considered normal or acceptable for modern AMD cards. I have not been able to find other reports of the same problem, so I'm assuming this is some sort of hardware fault. Do you think this is correct? To I need to put any other components on my suspect list when I begin replacing hardware?
The next thing I'm going to do is to buy and install an NVidia card. It feels like they are the only viable option for dedicated graphics right now. That makes me feel very uncomfortable. I don't want a monopoly, but it's pointless to put up with something broken just to provide a counterweight. Does anyone understand my dilemma?
-
today i tried putting on my ASUS 15.6-Inch X551MAV Notebook - Celeron Processor computer, it too about an hour to come on,afterwhich i put in the charger the and it is not coming on,in addition to that when i tried to access the router in my home,the connection is showing really low,and im frustrated because i have all my assignments on there which is due next week
-
Good luck -
Good luck -
Have you contacted the reseller? You will have warranty. That should be the first thing you do before opening up its guts. They can't help you with a problem you don't tell them about.
+1 on suspecting system RAM - reseat it and run a memtest. Let it run for hours. Or try removing all but one stick at a time and seeing if you still get crashy crash.
Radeons and their drivers don't have a great mobile GPU rep but they're not THAT broken that your symptoms are normal behavior out of the box... -
I haven't tried a memtest, I will try that overnight and see if anything turns up.
About warranty, I think I've already voided my warranty with BIOS flashing and changing the clock speeds... I haven't read the fine print, but I believe that is standard. I've also opened it up already to insert an SSD. The second problem with that is that I'll probably be without a laptop for a long time. I mean it's still better than nothing. -
Yeah, try memtest before trying anything else.
About warranty: changing HDD (or SSD) does not void warranty. Not sure BIOS flashing does, I would not think so. Changing clock speeds does I assume, esp. if you overclock, but you could flash the old one back ... if you have a backup.
Good luck
Recurring GPU driver crash with HD 8970M in a P170SM
Discussion in 'Other Manufacturers' started by Oddwarg, Oct 3, 2014.