Graphic Cards Products to not buy - MSI Radeon HD 7850 Power Edition

Hi everyone!

After a lot of fighting with MSI India and Taiwan, they finally agreed to switch my card for another one from their range. This was after they tested my card for over a month and concluded that there was nothing wrong with it to begin with. It was also stress-tested in Taiwan.

I have picked the GTX670 OC Edition. All problems that I was facing have vanished for good, though it did cost me around 10k more.

I still don't understand what exactly went wrong there but I'm glad to be back with Nvidia. The drivers are much more reliable and stable.

Also, I don't blame MSI's product quality. I blame AMD's shitty drivers. They have always got it all wrong with drivers!

If anyone is thinking about this card, I'd not recommend otherwise, but you should be aware that AMD still has to get their drivers straight. They're buggy and extremely unreliable!

With the same everything else, I'm super satisfied with the stability of my new GTX670..

And finally, the solution that I found for a friend with te same GPU and problem. It was some random conflict wit the RAM. I removed 2x4GB out of the 4x4GB and there were no more crashes. The RAM was a Corsair XMS3, which runs at 1.65V. You should go for a 1.5V RAM instead.
 
You're joking, right? I've been using AMD's drivers and with multiple cards over the years. My 6970 Crossfired cards run rock stable overclocked well beyond their limits and on water. Both vendors have been on and off with their drivers - and I've been using PCs since 1985, before ATI and nVidia were even born - and some forumites. nVidia is not the holy know-it-all either, I've owned enough of both brands and used their drivers extensively - or had to.

Stop spreading FUD. It is your personal experience and probably some issue with your system or setup if the card was all clear. Everything you experience is not gospel truth. Your RAM issue was obviously what was creating the conflict.
 
You're joking, right? I've been using AMD's drivers and with multiple cards over the years. My 6970 Crossfired cards run rock stable overclocked well beyond their limits and on water. Both vendors have been on and off with their drivers - and I've been using PCs since 1985, before ATI and nVidia were even born - and some forumites. nVidia is not the holy know-it-all either, I've owned enough of both brands and used their drivers extensively - or had to.

Stop spreading FUD. It is your personal experience and probably some issue with your system or setup if the card was all clear. Everything you experience is not gospel truth. Your RAM issue was obviously what was creating the conflict.

+1 with this.Sapphire 7970 owner here.NO problems faced for the last 4 months.
 
You're joking, right? I've been using AMD's drivers and with multiple cards over the years. My 6970 Crossfired cards run rock stable overclocked well beyond their limits and on water. Both vendors have been on and off with their drivers - and I've been using PCs since 1985, before ATI and nVidia were even born - and some forumites. nVidia is not the holy know-it-all either, I've owned enough of both brands and used their drivers extensively - or had to.

Stop spreading FUD. It is your personal experience and probably some issue with your system or setup if the card was all clear. Everything you experience is not gospel truth. Your RAM issue was obviously what was creating the conflict.

First off, I never prescribed it as the gospel truth. If anything, I said that if one is thinking about this card, I'd not recommend otherwise. These problems don't show up in Linux at all. You don't need to have 30 years of PC experience to deduce this.

Secondly, my RAM is perfectly fine and I can say this after countless hours of memtest on each module. The problem is when the AMD card is put in with the RAM. It doesn't persist with the Nvidia card. What would be your conclusion if you were in my place? Has ATI/ AMD ever documented this that the said RAM can cause crashes? Is it anywhere on their website?

I spent good money on my rig. I didn't cheap out on any component. Why should I tolerate any incompatibilities at all? Why should I buy the uncertainty that ships with that product? I don't think is is unwise to spend that extra 2-3k for your peace of mind. I wanted to like AMD this time. I was happy with the performance I got, for about 2 minutes before my system invariably froze each time.

This isn't an Nvidia vs AMD discussion. I have friends who swear by AMD/ ATI and their rigs are great. I have used ATI when 9700 PRO came out. It was awesome.

If you're saying that I'm spreading FUD here, I request you to find a logical explanation for the problem. My configuration is as follows:

GPU: MSI R7850 Power Edition 2GB GDDR5/ MSI GTX 670 2GB GDDR5
CPU: Intel Core i7 2600
Motherboard: ASUS P8Z68V PRO
RAM: Corsair XMS3 2 x 4GB RAM (1600MHz at 1.65V)
PSU: Corsair 650TX
Cooling: Corsair H80 CPU liquid cooling system
HDD: WD Black 1000GB + Seagate Barracuda 1000GB (not in RAID)
Optical: ASUS DVD+/-RW
(note: The only reason why you see so much of Corsair is that my local dealer only sells Corsair. And the fact that whatever they are selling is, to say the very least, passable.)
OS: Windows 8 Pro x64 on UEFI mode

Off-topic: The GTX 670 heats up a lot! On full load it runs at around 88 degrees Celsius and sometimes even higher. Nothing's perfect.. Sigh..
 
I guess you are right but this problem is due to the faulty card you got... It happens sometimes bro..and I know its really frustrating when you spend this much on a GPU and things go wrong....bw I have also heard problems with MSI 7850 power edition(but only & only with this particular card).. It might be coz of any reason but i was also warned by my dealer to go for another card coz msi 7850 power edition is having some problem as he received some of them back for rma.... But really amd and nvidia are equally good...
 
Off-topic: The GTX 670 heats up a lot! On full load it runs at around 88 degrees Celsius and sometimes even higher. Nothing's perfect.. Sigh..

Sorry to hear that! How did you monitor the temperatures?

What about the ventilation inside the chassis? Which chassis do you have ATM? And how are the fans inside the chassis configured (i.e. inlet/exhaust)? Insufficient ventilation inside the chassis can cause the internals to heat up quickly.
 
<snip>AMD's shitty drivers. They have always got it all wrong with drivers!<snip>

<snip>you should be aware that AMD still has to get their drivers straight. They're buggy and extremely unreliable!<snip>

<snip>It was some random conflict wit the RAM. I removed 2x4GB out of the 4x4GB and there were no more crashes. The RAM was a Corsair XMS3, which runs at 1.65V. You should go for a 1.5V RAM instead.<snip>

Your post was unequivocal in you blaming AMD's drivers, and that they have got it all wrong, and buggy and extremely unreliable, and yet to get it straight. Your words, not mine.

Also, *you* said it was a RAM issue, I didn't. The only common link between a video card and the memory is the CPU, because the CPU has both the controllers on its die. In real life the twain do not have many conversations directly, so compatibility should never be a problem except in your case where it was. Without trying out every possible permutation of RAM speed and timing, it would be difficult to point blame. And in this day and age, trouble is not avoidable even my spending lakhs of rupees. It is to be tackled appropriately when it does happen, and that is true for almost anything. Taking judgmental stands does not solve anything. Yes switching out components is one way to do it but then one never figures out the real root of the problem.
 
Your post was unequivocal in you blaming AMD's drivers, and that they have got it all wrong, and buggy and extremely unreliable, and yet to get it straight. Your words, not mine.

Also, *you* said it was a RAM issue, I didn't. The only common link between a video card and the memory is the CPU, because the CPU has both the controllers on its die. In real life the twain do not have many conversations directly, so compatibility should never be a problem except in your case where it was. Without trying out every possible permutation of RAM speed and timing, it would be difficult to point blame. And in this day and age, trouble is not avoidable even my spending lakhs of rupees. It is to be tackled appropriately when it does happen, and that is true for almost anything. Taking judgmental stands does not solve anything. Yes switching out components is one way to do it but then one never figures out the real root of the problem.


As I mentioned, the Linux driver was working fine, which was surprising as AMD/ ATI Linux drivers don't have a great reputation (right from the starting of the fglrx days). I was able to play some games without any problems at all. This indicated that the problem could possibly not have been with the hardware (yes, I did use all my 4 x 4GB), but with the software.

With Windows: My MSI Geforce 9600GT didn't have any problems, but the AMD/ ATI R7850 was still freezing.

This is all using the RAM manufacturer's specified/ recommended SPD settings.

If it wasn't the driver, what could it be?

The worst part was that the bugger worked perfectly at MSI's service center!

A friend of mine in Delhi has the same everything as mine (we bought it around the same time) down to the PSU and RAM. He uses a Radeon 7950 without any problems.

Honestly, I haven't seen anything like this before and since the drivers were the most obvious culprit, I spent each moment cursing and bashing them.. :p

For all you know it is the drivers with THIS particular card. I'll never know now that it's gone.

Sorry to hear that! How did you monitor the temperatures?

What about the ventilation inside the chassis? Which chassis do you have ATM? And how are the fans inside the chassis configured (i.e. inlet/exhaust)? Insufficient ventilation inside the chassis can cause the internals to heat up quickly.

The ventilation is fine. The main issue is that I'm in Kota, Rajasthan and this place is hot as hell. I can't even think of using boost-clocks. It is so hot that in spite of using the H80, my CPU still peaks at around 60 degrees at full load.
 
That's a fair way of looking at it. It is quite possible your card really had an issue, and only the windows driver exposed it. Stranger things have been known to happen. Windows and linux have different ways of addressing drivers in memory, so an underlying problem may be masked in either os.

My point and subsequent beef was that MSI has not been receiving every card like yours back, and that there is a massive installed base of amd cards, so bashing the driver and the card was not on. Compatibility and hardware faults are part of life and that is what warranties are for.

Look if your problems went away when you removed a single stick of memory, then the driver was obviously not the culprit. The driver has no way of knowing how much and what ram you have. It doesn't even care.

I can only think of the cpu, memory and card as culprits. In any case you should run basic tests to eliminate the cpu and memory, specially the memory, to resolve any remaining doubts.
 
Lesson Learnt: Buy memory with standard voltage requirements. 'Exotic' voltages will cause problems.

At least MSI was kind enough to swap it for me. They had every right to not entertain my request after the card was found ok. The guy at the service center later told me that out of the thousands of cards sold, he was only aware of two cases where this card had caused problems.
 
They may be good. But I've seen them cause problems in at least 3-4 more computers.

If budget allows it one should go for their Vengeance series. G-Skill is even better but there is a support problem in smaller towns.
 
Back
Top