How many of you Radeon 5700 owners have ditched your cards over the drivers?

[EarthDog]

OK. That's an isolated driver issue with that particular card.

My point was that both Nvidia and AMD have various driver issues. In the past and present and I'm sure in the future too. But people are making it look like AMD drivers are bad and Nvidia's are like little Angels lol clearly not the case.

Did anyone do that, though, really?

Neither are angels, indeed, but let's be honest....right now and at least since the release of Navi, but especially since the release of Adrenalin 2020, AMD drivers are not as stable as Nvidia's. The fix list since A2020 has been long and the fixes frequent (the latter certainly a GOOD thing). AMD even admitted they had an issue with the A2020 driver!

If people are saying Nvidia is problem free, they are just as much delusional as some are denying AMD currently (and for the last several months) has more issues. They are working their arse off to fix it as we can see with hotfixes and the unusually close releases... so any that still have issues, it's a WIP.

 

[Super XP]

Did anyone do that, though, really?

Neither are angels, indeed, but let's be honest....right now and at least since the release of Navi, but especially since the release of Adrenalin 2020, AMD drivers are not as stable as Nvidia's. The fix list since A2020 has been long and the fixes frequent (the latter certainly a GOOD thing). AMD even admitted they had an issue with the A2020 driver!

If people are saying Nvidia is problem free, they are just as much delusional as some are denying AMD currently (and for the last several months) has more issues. They are working their arse off to fix it as we can see with hotfixes and the unusually close releases... so any that still have issues, it's a WIP.

YES.
But I suspect Nvidia Fanboys that don't own Radeon cards are regurgitating actual Radeon owner complaints.

Correct there is some issues with AMD AD2020 drivers but at the same time a lot of the issues could have been mitigated with proper uninstall and reinstall. User error is another issue, as stated a few times on Reddit. People installed new 5700XTs without resetting there Overclocked memory and CPUs. As soon as they reset those memory overclocks no more 5700XT stability issues. Then they work there Ram speed back up gradually.

 

[Vya Domus]

Hardware Unboxed said AMD RMA is 5x higher than Nvidia. Where are official Mindfactory.de RMA numbers? Not some alternative Reddit numbers?

Imagine the sort of mental gymnastics you'd have to pull to believe that numbers that come from an actual retailer are less believeble than the ones that come from something like Hardware Unboxed.

The numbers are made up, don't worry mate, carry on.

 

[Assimilator]

YES.
But I suspect Nvidia Fanboys that don't own Radeon cards are regurgitating actual Radeon owner complaints.

As opposed to you, a Radeon fanboy who just s**tposts in any thread critical of AMD for whatever reason.

Here's a novel idea, Captain Genius: us NVIDIA card owners want to buy AMD cards, but aren't willing to put up with driver issues. We want AMD to fix its drivers so we can be comfortable buying their GPUs. We want to give AMD our money, but aren't prepared to put up with substandard driver quality.

Correct there is some issues with AMD AD2020 drivers but at the same time a lot of the issues could have been mitigated with proper uninstall and reinstall. User error is another issue, as stated a few times on Reddit. People installed new 5700XTs without resetting there Overclocked memory and CPUs. As soon as they reset those memory overclocks no more 5700XT stability issues. Then they work there Ram speed back up gradually.

And the majority of people with AMD GPU driver issues don't have overclocks, or dialling their OCs down doesn't make any difference. So did you have a point, or are you just here to spread whataboutism as usual?

 

[Super XP]

As opposed to you, a Radeon fanboy who just s**tposts in any thread critical of AMD for whatever reason.

Here's a novel idea, Captain Genius: us NVIDIA card owners want to buy AMD cards, but aren't willing to put up with driver issues. We want AMD to fix its drivers so we can be comfortable buying their GPUs. We want to give AMD our money, but aren't prepared to put up with substandard driver quality.



And the majority of people with AMD GPU driver issues don't have overclocks, or dialling their OCs down doesn't make any difference. So did you have a point, or are you just here to spread whataboutism as usual?

I already made a valid and substantiated point with regards to Ram overclocks causing issues. I also agreed many people are having legitimate driver issues too.

Claiming otherwise is calling a bunch of people on Amazon, Newegg and Reddit reviews and comments a bunch of imbiciles and liars. ARE They??

That's interesting, as I choose Radeon graphics for stability as I've had issues with Nvidia drivers in the past. Though I'll probably never purchase a GeForce card for the foreseeable future Only because of what Nvidia stands for. Deep throating the industry with its GPP scam. But that's a different discussion on a different thread.

 

[freeagent]

I would buy one, but not with these driver issues. I come from a land where everything works, I’m not willing to drop good money on good hardware with broken software. I would buy if they can fix their software, which sounds like they are, slowly but surely. There is hope.

 

[Darmok N Jalad]

I’ve played three different titles for several hours now. I had one crash, but I believe that was due to too much of an undervolt. It’s otherwise no-nonsense work. I wonder how many returns are simply due to the blower on early models. It does have the ability to get loud, and that might have been off-putting for many.

 

[Super XP]

I’ve played three different titles for several hours now. I had one crash, but I believe that was due to too much of an undervolt. It’s otherwise no-nonsense work. I wonder how many returns are simply due to the blower on early models. It does have the ability to get loud, and that might have been off-putting for many.

I hope AMD never releases a Blower Reference GPU again. The Radeon VII cooler should be the standard reference cooler for Radeon cards. It cools great and looks great.

 

[bug]

I’ve played three different titles for several hours now. I had one crash, but I believe that was due to too much of an undervolt. It’s otherwise no-nonsense work. I wonder how many returns are simply due to the blower on early models. It does have the ability to get loud, and that might have been off-putting for many.

Does it really matter what's the cause? At the end of the day a disappointed user is a disappointed user. Not only are they very likely to avoid the make for a few generations, but they'll also give you a good deal of badmouth (sp?).

 

[Vayra86]

Ironically the 2080ti still has the highest failure rate and those are supposed to be the pinnacle of premium cards.

Its not entirely suprising for such a massive die and higher TDP card to have higher return rates. Lower total number and less options for binning any failures... Falls right in line with most other big chips over the years I think. Also remember the Space Invaders bad batch. That's definitely made a dent. And even today I do see that 2080ti topic pop up every so often, with a problem card. Much more so than any other Turing gpu except for Gigabyte's blocks of horror.

What's much more surprising to me is that Navi's are at least double the return rate of similar performance Turing cards, while those Turing dies are much bigger, too. And that is even discounting the outliers which consist mainly of very weak AIB releases (10%?! holy shit... its a miracle Powercolor is still playing the game). Now thát is not unique to AMD. But those Navi returns are high across the board, or at least, notably higher than the competitor.

Premium or not has never had any sort of effect on return rates, quite the contrary. But if a mass produced midranger has higher return rates, that is very costly. You're not talking about a few dozen cards now, but a few thousand, and on top of that, the margins are not quite as high per card either.

YES.
But I suspect Nvidia Fanboys that don't own Radeon cards are regurgitating actual Radeon owner complaints.

Correct there is some issues with AMD AD2020 drivers but at the same time a lot of the issues could have been mitigated with proper uninstall and reinstall. User error is another issue, as stated a few times on Reddit. People installed new 5700XTs without resetting there Overclocked memory and CPUs. As soon as they reset those memory overclocks no more 5700XT stability issues. Then they work there Ram speed back up gradually.

Ehh gonna stop you there.

User error? Let's go over a little comparison if you're not a total geek like one of us but a random Joe wanting to play games.
- Nvidia cards don't require DDU, I've never ever used it even for an AMD> NV switch... In every single case it was plug and play, really. Beyond that... overclocking was two clicks and done since Kepler. The OC was stable and if it wasn't it would show you within minutes of play. Temps, even on the blower, were kept in check because of a strict driver/BIOS regime. Yes, it would throttle. But it wouldn't hit 90C and overall performance remained normal.
- Can't install Navi without resetting memory and CPU OC? How is that normal? Especially for Joe, how the hell would he know? Working up to recover your old OC because a new GPU was inserted? What madness is this? And what does that balance trick say about the general stability of your GPU?

See, that is the difference in a nutshell right there. An Nvidia card is fire and forget. An AMD card is like UT'99's Redeemer, if you don't keep tight control, you're not gonna have a good time. And its a different sort of beast every. single. time. If its not BIOS, or post launch spec changes, or shitty AIB versions, its the actual drivers from AMD itself. And if all is well, the cards aren't in stock...

Does it really matter what's the cause? At the end of the day a disappointed user is a disappointed user. Not only are they very likely to avoid the make for a few generations, but they'll also give you a good deal of badmouth (sp?).

This, precisely. Its all about consistency, and faith in company/product. Hard to build, easy to lose.

 

[ratirt]

Funny thing is i never had a driver issue but I've encountered hardware issue (return the card, NV card to be exact) Had cards from both camps and never had an issue with the driver. Sometimes tweaks were in order or reinstalling but that is mostly it. My buddies have no issues either and they've got 5700 and 5700 XT's 5600XT you name it. I'm really surprised you are still on this bullshit. Game can have an issue with the driver and needs some tweaks. You can run apps accelerated by GPUs and no issues. Run a game and something isn't right but with this one all other 10 are OK.
We need to understand few things.
1st it is something new and it may happen that there will be issues.
2nd it may be a hardware issue not related to the chip or the driver but the gpu producer. (asus or xfx)
3rd user laziness and lack of basic knowledge (and in my book this is a 90% of the issues). I always had a saying when dealing with users in the past when being an administrator and help users overcome their problems. Stop user'ing, give a detailed explanation what the problem is an maybe think what you have done wrong. It is mostly human factor that something craps out.
How many times someone did something stupid blaming hardware or software. There was only one major problem. Person that is using it.

 

[Vayra86]

Funny thing is i never had a driver issue but I've encountered hardware issue (return the card, NV card to be exact) Had cards from both camps and never had an issue with the driver. Sometimes tweaks were in order or reinstalling but that is mostly it. My buddies have no issues either and they've got 5700 and 5700 XT's 5600XT you name it. I'm really surprised you are still on this bullshit. Game can have an issue with the driver and needs some tweaks. You can run apps accelerated by GPUs and no issues. Run a game and something isn't right but with this one all other 10 are OK.
We need to understand few things.
1st it is something new and it may happen that there will be issues.
2nd it may be a hardware issue not related to the chip or the driver but the gpu producer. (asus or xfx)
3rd user laziness and lack of basic knowledge (and in my book this is a 90% of the issues). I always had a saying when dealing with users in the past when being an administrator and help users overcome their problems. Stop user'ing, give a detailed explanation what the problem is an maybe think what you have done wrong. It is mostly human factor that something craps out.
How many times someone did something stupid blaming hardware or software. There was only one major problem. Person that is using it.

TL DR 'I never had an issue so it doesn't exist and everyone is full of it'

1st its not new, the card is seven months old
2nd it may be anything but does it matter? Product doesn't work as advertised.
3rd See post above.

Blaming user error here is so far off the mark... unreal

 

[ratirt]

TL DR 'I never had an issue so it doesn't exist and everyone is full of it'

Blaming user error here is so far off the mark... unreal

yes I haven't had an issue. (I had but I was able to deal with it myself). It is not to say there is no issues because I haven't had any.
not saying it may not exist but I've seen it. experienced it and I know how people are reacting. It is way better to blame hardware or a driver than incompetency of their own.

Why is it unreal? We have all used same drivers and I haven't seen or been told by my friends having any problems? If it were a driver issue the problems would happen on all the 5000 series cards wouldn't they? One game with a driver problem is not driver issue. At least for me. My friends say no driver problems and I tend to believe them cause they know what they are doing.

 

[Vayra86]

yes I haven't had an issue. (I had but I was able to deal with it myself). It is not to say there is no issues because I haven't had any.
not saying it may not exist but I've seen it. experienced it and I know how people are reacting. It is way better to blame hardware or a driver than incompetency of their own.

Why is it unreal? We have all used same drivers and I haven't seen or been told by my friends having any problems? If it were a driver issue the problems would happen on all the 5000 series cards wouldn't they? One game with a driver problem is not driver issue. At least for me. My friends say no driver problems and I tend to believe them cause they know what they are doing.

We used the same drivers... so that means all variables are accounted for? That is not sound logic in a PC environment... you know this. There are many, many configurations, OS versions, etc etc. And everything is in motion too, people upgrade and update whatever in whichever order. The conclusion there should be that a driver should be rock solid and capable of doing its work despite all of that stuff. Not that it 'might fail'... last I checked, vendors don't put a sticker on the box saying 'driver might fail'

One game with a driver problem, is one game with a driver problem that warrants a fix, either through the game or the driver. Simple. But that isn't the issue here, hell I've yet to see any ingame problems, all I see is general, game breaking issues that never even get you to a launcher.

 

[Vya Domus]

Its not entirely suprising for such a massive die and higher TDP card to have higher return rates. Lower total number and less options for binning any failures... Falls right in line with most other big chips over the years I think. Also remember the Space Invaders bad batch. That's definitely made a dent. And even today I do see that 2080ti topic pop up every so often, with a problem card. Much more so than any other Turing gpu except for Gigabyte's blocks of horror.

You're making the assumption that it's the actual chips which cause the problems but they are clearly not, not in the case of Navi nor RTX 2080ti. If that was the cause you'd see a consistent fail rate among all cards but clearly it depends on the brand and model. In other words the AIB partners do a horrid job, that's why I am amazed you can pay 1000$ and still get a product that is of no better quality than if you paid 300$.

 

[ratirt]

We used the same drivers... so that means all variables are accounted for? That is not sound logic in a PC environment... you know this. There are many, many configurations, OS versions, etc etc. And everything is in motion too, people upgrade and update whatever in whichever order. The conclusion there should be that a driver should be rock solid and capable of doing its work despite all of that stuff. Not that it 'might fail'... last I checked, vendors don't put a sticker on the box saying 'driver might fail'

One game with a driver problem, is one game with a driver problem that warrants a fix, either through the game or the driver. Simple. But that isn't the issue here, hell I've yet to see any ingame problems, all I see is general, game breaking issues that never even get you to a launcher.

Doesn't it sound logically ? Because for me it does tell something. Same driver different issues what is that tell you? I agree there are many configurations. That is obvious. What I'm trying to say you cannot generalize something to one variable where there's 100 of them and you have already admitted it. Would you agree? For me the driver is rock solid. Here's an example which I've come across few weeks ago.
You tell me what was the issue here.
Recently I stepped on the CS:GO bandwagon and I been playing for weeks now. With my Vega 64 I got everything sorted with the settings and played for a week. I never use automatic updates for anything so as you can presume the graphics driver auto-update is off. One day I got a CS:GO game update. (as you know, updates are being released systematically not just for graphics driver but for everything) Strange thing happened, while playing, my screen went black. Just like that. cold reset only solution. Started up the game again. Same thing but played a while longer maybe. Looked over system events. Said graphics driver crashed. Well tried tweaking but it didn't help so restored to my previous settings and called it a day. 3 days later another update for the game. Problem gone and I didn't reinstall any graphics driver nor updated the graphics driver or reinstalled the game. I simply played something else. Now, is this a driver issue? The win event log said graphics driver crash so it would seem so but it didn't say that the game caused the driver to crash and that was the problem which was gone after another update 3 days later. That is just one of many examples.

Since you mentioned that there is so many variables that may cause the application to crash or the driver, I'm really surprised that you have generalized it to only graphics driver issues. (it is simple to blame this only) Imagine that 99% of users would return their cards (and I hear about this in forums the black screen of death) because of this saying graphics driver problem. For me this is total bullshit and we know how many new games have updates weekly especially at their release. There can be driver issues. We have seen the updates recently. There's always something you can improve not necessarily issues but performance or add more features. I will look at the list of last 4 AMD driver updates to make sure what was fixed and what the drivers actually improve but I already know the answer. Non of us actually knows what happened with individual user's problem for their card to be returned blaming drivers and you already know why. He will say driver issue but was it really? Knowing how these users operate which I have mentioned earlier I really doubt it was a driver issue. It was an issue though. That is my logic take it from this what you will. I could go on about this and believe me the real graphics driver issue (if there was any major graphics driver problem) has been fixed few months ago.

 

[btarunr]

Before the RTX 2080 Super in my machine, I had a reference RX 5700 XT. Before Adrenalin 2020, I think till 19.10, the drivers were so bad that I wanted to yank the card out and yeet it. All my problems disappeared with 19.10. Then in December I got this RTX 2080 Super, and retired the RX 5700 XT to another machine that I use at least 2-3 hours a day, including BFV. Surprisingly, no problems with Adrenalin 2020.

 

[Vayra86]

Doesn't it sound logically ? Because for me it does tell something. Same driver different issues what is that tell you? I agree there are many configurations. That is obvious. What I'm trying to say you cannot generalize something to one variable where there's 100 of them and you have already admitted it. Would you agree? For me the driver is rock solid. Here's an example which I've come across few weeks ago.
You tell me what was the issue here.
Recently I stepped on the CS:GO bandwagon and I been playing for weeks now. With my Vega 64 I got everything sorted with the settings and played for a week. I never use automatic updates for anything so as you can presume the graphics driver auto-update is off. One day I got a CS:GO game update. (as you know, updates are being released systematically not just for graphics driver but for everything) Strange thing happened, while playing, my screen went black. Just like that. cold reset only solution. Started up the game again. Same thing but played a while longer maybe. Looked over system events. Said graphics driver crashed. Well tried tweaking but it didn't help so restored to my previous settings and called it a day. 3 days later another update for the game. Problem gone and I didn't reinstall any graphics driver nor updated the graphics driver or reinstalled the game. I simply played something else. Now, is this a driver issue? The win event log said graphics driver crash so it would seem so but it didn't say that the game caused the driver to crash and that was the problem which was gone after another update 3 days later. That is just one of many examples.

Since you mentioned that there is so many variables that may cause the application to crash or the driver, I'm really surprised that you have generalized it to only graphics driver issues. (it is simple to blame this only) Imagine that 99% of users would return their cards (and I hear about this in forums the black screen of death) because of this saying graphics driver problem. For me this is total bullshit and we know how many new games have updates weekly especially at their release. There can be driver issues. We have seen the updates recently. There's always something you can improve not necessarily issues but performance or add more features. I will look at the list of last 4 AMD driver updates to make sure what was fixed and what the drivers actually improve but I already know the answer. Non of us actually knows what happened with individual user's problem for their card to be returned blaming drivers and you already know why. He will say driver issue but was it really? Knowing how these users operate which I have mentioned earlier I really doubt it was a driver issue. It was an issue though. That is my logic take it from this what you will. I could go on about this and believe me the real graphics driver issue (if there was any major graphics driver problem) has been fixed few months ago.

The whole god damn point of a good driver is that it can withstand all the movement in other areas of a system or its software. And also withstand differences in configuration, such as an overclock.

And in a general sense, AMD screws that up more often than it does get things right. And when they get it right, they add new features to kill stability once more. By comparison, if you get a new game release that somehow doesn't work properly on an Nvidia card, there is a Game Ready driver to fix that, and if that still has problems, there is a hotfix to follow up shortly after. Thát is how these things should be working, and with AMD they generally do not. That is also, I believe, what we as customers should be expecting from our GPU vendors, regardless of whether its red or green.

In the same way I expect game devs to do their job. If a game has a bug and its clearly not a GPU problem, I expect a patch. Yesterday, really. After all, how did that bug pass testing? What does it say about a company's internal processes. In my line of work... I should try bringing new stuff to a production environment untested for once... I can pack my bags a day after.

In the end its about the big data behind it. The numbers, not the individuals, and the trends, not the individual events. And the trend is clear.

You're making the assumption that it's the actual chips which cause the problems but they are clearly not, not in the case of Navi nor RTX 2080ti. If that was the cause you'd see a consistent fail rate among all cards but clearly it depends on the brand and model. In other words the AIB partners do a horrid job, that's why I am amazed you can pay 1000$ and still get a product that is of no better quality than if you paid 300$.

Absolutely true, but the balance is still tighter with a higher TDP chip and board. Its definitely easier, and also cheaper, to make a smaller GPU. The chip is never the problem but it still is a binned part of the GPU. And the lower down the stack you go the more options you have to bin correctly. AMD had a similar example with that 14Gbps VRAM thing. This is also a binning issue at its core, and a vague attempt to make it look like lower binned parts can perform as higher ones when in fact only a small portion of them is capable. Another example is the consistent story about undervolting AMD cards. Why do you think this is a thing? Its not because AMD likes to advertise higher TDPs... its because they cannot provide assurance of stability with lower TDPs. It tells us that their chips don't come out of the fab as consistently as they should, or could be.

Its a many headed beast I think, and every player is trying to cut cost. AIBs do that by placing weak components on boards and re-using old cooler designs. If it can pass a test and not burn down, its probably good enough...

 

[Super XP]

Its not entirely suprising for such a massive die and higher TDP card to have higher return rates. Lower total number and less options for binning any failures... Falls right in line with most other big chips over the years I think. Also remember the Space Invaders bad batch. That's definitely made a dent. And even today I do see that 2080ti topic pop up every so often, with a problem card. Much more so than any other Turing gpu except for Gigabyte's blocks of horror.

What's much more surprising to me is that Navi's are at least double the return rate of similar performance Turing cards, while those Turing dies are much bigger, too. And that is even discounting the outliers which consist mainly of very weak AIB releases (10%?! holy shit... its a miracle Powercolor is still playing the game). Now thát is not unique to AMD. But those Navi returns are high across the board, or at least, notably higher than the competitor.

Premium or not has never had any sort of effect on return rates, quite the contrary. But if a mass produced midranger has higher return rates, that is very costly. You're not talking about a few dozen cards now, but a few thousand, and on top of that, the margins are not quite as high per card either.



Ehh gonna stop you there.

User error? Let's go over a little comparison if you're not a total geek like one of us but a random Joe wanting to play games.
- Nvidia cards don't require DDU, I've never ever used it even for an AMD> NV switch... In every single case it was plug and play, really. Beyond that... overclocking was two clicks and done since Kepler. The OC was stable and if it wasn't it would show you within minutes of play. Temps, even on the blower, were kept in check because of a strict driver/BIOS regime. Yes, it would throttle. But it wouldn't hit 90C and overall performance remained normal.
- Can't install Navi without resetting memory and CPU OC? How is that normal? Especially for Joe, how the hell would he know? Working up to recover your old OC because a new GPU was inserted? What madness is this? And what does that balance trick say about the general stability of your GPU?

See, that is the difference in a nutshell right there. An Nvidia card is fire and forget. An AMD card is like UT'99's Redeemer, if you don't keep tight control, you're not gonna have a good time. And its a different sort of beast every. single. time. If its not BIOS, or post launch spec changes, or shitty AIB versions, its the actual drivers from AMD itself. And if all is well, the cards aren't in stock...



This, precisely. Its all about consistency, and faith in company/product. Hard to build, easy to lose.

I think you completely miss understood what I was trying to explain. And this isn't me experimenting, this is several comments from others off Reddit and forums. If you have overclocked UNSTABLE MEMORY or CPU, then install a Radeon Graphics Card, and you get Blue Screens, how is that the GPU's fault?? The suggestion was to default stock your CPU & RAM, install your graphics cards, only to Eliminate an ALREADY unstable CPU & RAM overclock, so they wouldn't blame the GPU, then once you have your GPU Drivers installed, go at it with the CPU & RAM OC if you choose.

Yes I'm a total geek & can assemble a gaming PC with my eyes closed if need be. lol

Oh and a FYI, there was plenty of Nvidia drivers that bricked or fried GPUs almost every year, some even pulled out of circulation by Nvidia.

 

[ratirt]

And also withstand differences in configuration, such as an overclock.

Each graphics card is different reacts differently with an OC not a driver problem. Still it will show up as a driver issue in the logs

The whole god damn point of a good driver is that it can withstand all the movement in other areas of a system or its software.

The whole system is complex, many variables as you said. Updates happen frequently. On one it works on the other it crashes. Even log will always say application crashed due to a driver issue. Not saying what caused the driver to crash. it may be not a driver issue but system has a problem. You can't expect a driver to fix system's bad config or a problem in the newly released driver.
Example: Look at the problems with the 1903 (i suppose it was that one for windows 10). (CPU usage on certain machines surges to 30%) Is that CPU's driver problem?

And in a general sense, AMD screws that up more often than it does get things right. And when they get it right, they add new features to kill stability once more. By comparison, if you get a new game release that somehow doesn't work properly on an Nvidia card, there is a Game Ready driver to fix that, and if that still has problems, there is a hotfix to follow up shortly after. Thát is how these things should be working, and with AMD they generally do not. That is also, I believe, what we as customers should be expecting from our GPU vendors, regardless of whether its red or green.

In general sense you want to say driver issue and that AMD screws that up because it is easy. You have no data about all the individual incidents and it is easier to blame AMD for driver issues.
Feature issues newly implemented in some games may occur (driver updates for AMD adrenaline) switch the feature off no driver issues. The feature needs tweaking because it is damn new and that is why.

In the same way I expect game devs to do their job. If a game has a bug and its clearly not a GPU problem, I expect a patch. Yesterday, really. After all, how did that bug pass testing? What does it say about a company's internal processes. In my line of work... I should try bringing new stuff to a production environment untested for once... I can pack my bags a day after.

Games problem. Windows event log will still say application crash due to a driver issue. Not a driver issue but still application caused the driver to crash.

I hope you see where I'm going with this. The actual driver issues are rare but it will always show up in the logs as a driver issue. Your perspective changes when you do actually realize, like you said, that there's 100 (or more) variables that can cause the issue but the driver will be the one to blame always. (or at least most cases). Graphics drivers are not for fixing other variables' wrong doings or fails but the driver is to display pictures.

In the end its about the big data behind it. The numbers, not the individuals, and the trends, not the individual events. And the trend is clear.

The trend is that 90% of users have no idea what they are doing and that trend my friend is growing. It is always easier to blame something else. So I disagree. Each individual incident is different. Numbers OK but the generalization of a problem is an issue here. Individual events reveal the actual problem. Not the number which are based on the cards returns because a user says it is a driver issues when they aren't actually.
It is wrong to call a monster truck a car but it is totally wrong when you call it a suspension bride. You my friend call all the returns of cards due to a driver issues or the problems that users are having are due to driver issues and that is totally wrong.

 

[Vayra86]

Each graphics card is different reacts differently with an OC not a driver problem. Still it will show up as a driver issue in the logs

The whole system is complex, many variables as you said. Updates happen frequently. On one it works on the other it crashes. Even log will always say application crashed due to a driver issue. Not saying what caused the driver to crash. it may be not a driver issue but system has a problem. You can't expect a driver to fix system's bad config or a problem in the newly released driver.
Example: Look at the problems with the 1903 (i suppose it was that one for windows 10). (CPU usage on certain machines surges to 30%) Is that CPU's driver problem?


In general sense you want to say driver issue and that AMD screws that up because it is easy. You have no data about all the individual incidents and it is easier to blame AMD for driver issues.
Feature issues newly implemented in some games may occur (driver updates for AMD adrenaline) switch the feature off no driver issues. The feature needs tweaking because it is damn new and that is why.

Games problem. Windows event log will still say application crash due to a driver issue. Not a driver issue but still application caused the driver to crash.

I hope you see where I'm going with this. The actual driver issues are rare but it will always show up in the logs as a driver issue. Your perspective changes when you do actually realize, like you said, that there's 100 (or more) variables that can cause the issue but the driver will be the one to blame always. (or at least most cases). Graphics drivers are not for fixing other variables' wrong doings or fails but the driver is to display pictures.

There is an abstraction layer, or several, between driver/GPU and the rest of the system for all the data it gets. There goes your story... the GPU is supposed to feed off that, and not all other sorts of requirements in the system. In the end, thát is really what is happening here, and drivers are really an assortment of compatibility bits to cater for a wide range of situations. The quality of the driver is expressed like that: how compatible is it, to all those situations. How does it handle them in case of a problem. Will it CTD, will it give you a TDR (Timeout/Device removed; as is Nvidia's way of keeping the system operational with the GPU in safe mode), or will it already guide the user to recommended / restricted settings. Or will you get a BSOD with some random error string.

So yes, I get you, but I still disagree we should not blame the driver The driver is the most fundamental piece of the puzzle here, for all software running on a GPU.

After that it is up to Nv/AMD to fight the battle with devs over who should change what. NOT my problem and once it becomes one, said company is on my shitlist. When Nvidia did not bring SLI support for Elder Scrolls Online back in the day, it was clear as day to me: this is the last SLI setup ever, go screw yourselves. Trust is hard to gain, and instantly lost. And in hindsight, that was a very good decision, Look where multi GPU is at today.

That is exercising customer power. Act on your findings and support quality. Being lenient on companies for screwing up only allows them to repeat it.

I think you completely miss understood what I was trying to explain. And this isn't me experimenting, this is several comments from others off Reddit and forums. If you have overclocked UNSTABLE MEMORY or CPU, then install a Radeon Graphics Card, and you get Blue Screens, how is that the GPU's fault?? The suggestion was to default stock your CPU & RAM, install your graphics cards, only to Eliminate an ALREADY unstable CPU & RAM overclock, so they wouldn't blame the GPU, then once you have your GPU Drivers installed, go at it with the CPU & RAM OC if you choose.

Yes I'm a total geek & can assemble a gaming PC with my eyes closed if need be. lol

Oh and a FYI, there was plenty of Nvidia drivers that bricked or fried GPUs almost every year, some even pulled out of circulation by Nvidia.

Point taken, and great, so let's cross that one off then. But it does seem odd that the insertion of a GPU is cause to make a reddit topic about BSODs due to your non GPU OC, is it not? But if that was the case, how was it even relevant to begin with? Its clear that such a thing is not a GPU issue, so why even use it as an example...

I mean yes, I fully understand the point you and @ratirt are making here that not all problems originate from the GPU. But in every single case, the truth comes to light and the trend is that the vendor is very, very slow and rather weak at providing fixes for the problems that DO originate from the GPU. And the fixes are also not always good ones. But the biggest question mark to me is why AMD is still using a waterfall principle for its releases of drivers. Big releases instead of cutting them up in many smaller ones, working iteratively... definitely something Nvidia approaches in a much more modern way.
In case you're like WTF is he on about...:

What's the Difference? Agile vs Scrum vs Waterfall vs Kanban

What's the difference between Agile vs Scrum vs Waterfall vs Kanban? Here's everything you need to know about these project management methodologies.

www.smartsheet.com

And ehm... you're gonna have to back up that 'almost every year' of bricked GPUs due to drivers with some sources, because as far as I know, there has been one truly documented instance of it, maybe two.

 

[ratirt]

So yes, I get you, but I still disagree we should not blame the driver The driver is the most fundamental piece of the puzzle here, for all software running on a GPU.

ohh I totally disagree with that. If the CPU is a brain of the computer and graphics is it's eyes then analogically to the software, windows is a brain and graphics driver is its eyes. i hope you understand what I'm trying to say.

There is an abstraction layer, or several, between driver/GPU and the rest of the system for all the data it gets. There goes your story... the GPU is supposed to feed off that, and not all other sorts of requirements in the system. In the end, thát is really what is happening here, and drivers are really an assortment of compatibility bits to cater for a wide range of situations. The quality of the driver is expressed like that: how compatible is it, to all those situations. How does it handle them in case of a problem. Will it CTD, will it give you a TDR (Timeout/Device removed; as is Nvidia's way of keeping the system operational with the GPU in safe mode), or will it already guide the user to recommended / restricted settings.

Drivers are for displaying pictures not fixing system or other problems. The GPU doesn't feed off any other resources but it is not responsible for it's proper functionality like you say it should giving as an example "solid driver" statement.

After that it is up to Nv/AMD to fight the battle with devs over who should change what. NOT my problem and once it becomes one, said company is on my shitlist. When Nvidia did not bring SLI support for Elder Scrolls Online back in the day, it was clear as day to me: this is the last SLI setup ever, go screw yourselves. Trust is hard to gain, and instantly lost.

Never said it is your problem. we may have to face the problems that come but saying that AMD and NV is responsible for fixing something only related to a driver (just like many other resources) is not adequate to what it really is. I gave you a CS:GO example. That is not a driver issue. As always there's people coding these variables or resources needed for the graphics driver to do what it is supposed to do. If they make a mistake the driver will crash but the application is to blame. NV has had also these "driver issues" related to application crashes. Don't own a NV graphics but it may not be driver issue but application that cased the driver to crash.
If you think I'm wrong. Try it yourself. OC the card beyond it's capabilities. The game will crash and look it up in the windows Event log. It will say driver issue but is it really a driver issue?

 

[Vayra86]

If you think I'm wrong. Try it yourself. OC the card beyond it's capabilities. The game will crash and look it up in the windows Event log. It will say driver issue but is it really a driver issue?

Dude, this goes way deeper than that, and yes, like I've acknowledged a half dozen times now, we get it, other things are reported as a driver crash. No shit. Let's stop going in circles

ohh I totally disagree with that. If the CPU is a brain of the computer and graphics is it's eyes then analogically to the software, windows is a brain and graphics driver is its eyes. i hope you understand what I'm trying to say.

WHAT? This kinda underlines you have no idea what you're talking about. Windows is a brain and the graphics driver its eyes?! What are you smoking? Its just code being processed and handed over. Make mistakes and it crashes, simple enough.

 

[ratirt]

Dude, this goes way deeper than that, and yes, like I've acknowledged a half dozen times now, we get it, other things are reported as a driver crash. No shit. Let's stop going in circles

Of course it does go deeper. An ocean of possibilities and things that can go wrong. I'm just trying to tell you about inexperienced users and what they will say about an event or a problem. it will always be a driver issue and we can see it clearly now. It is not necessarily that. That is all and the numbers you mentioned earlier are not putting this aspect into consideration when spitting out a value.

 

[Darmok N Jalad]

Does it really matter what's the cause? At the end of the day a disappointed user is a disappointed user. Not only are they very likely to avoid the make for a few generations, but they'll also give you a good deal of badmouth (sp?).

Of course the cause matters. If AMD used a bad reference cooler, that is easy for a non-biased customer to make a judgement on the issue—“it’s a fast card, but it’s way too loud.” Having constant crashes and/or hardware failures is a much bigger issue—“this card just doesn’t work well at all.” It also leads to how experts can advise potential buyers—“should I buy a 5700?” The answer would either be “yes, but get a non-blower model,” or ”no, the card is unreliable.” We have discussions like this all the time where you may just need to avoid a particular brand. The discussion is about the reliability of this card, and so far my early conclusion is, just don’t get the blower version if noise is a concern.