[RESOLVED] AMD Driver reset with P3d V3

[mcp5500]

Unbelievable

[mcp5500]

It looks to me that you are making a statement that you will no longer support AMD.

[Optimist]

@mcp5500
Well, I can understand your anger in regards to this, and I am not all too happy about this, either. But trolling around won´t help anyone and this is just what your last three posts were about, sorry.

However, I do not see why AMD is to blame if software developers are using outdated drivers for their testing. On the one hand it is pronounced over and over again how V3 will take advantage of newer high end cards, but this is tested with a driver version in which the newer cards are not even officially supported.

Maybe it would have been a good idea to give some V2 users the possibility in participating in a closed Beta and offer V3 as a discount for their efforts, instead of selling it out to the public and now hoping that customers will spend their time voluntarily to final test their poduct.

Actually V3 is my first version, and I came here for getting a stable simulator with support and in return support a team still working actively on simulators like this.

At the moment, when reading through different posts, I am not so sure anymore how a lot of these mentioned issues could get into a final build. I am aware that all major releases have their flaws and bugs, but at some elemental level it should work, and there still is a difference of minor/major defects and a complete showstopper, which this AMD issue clearly is.

But ok, V3 is out for just one week, and let´s give the developers some time to debug.

In worst case, if there is no progress the next weeks and this program still cannot not even be started or used as offered, there is still the option to officially request a refund as sad as this might sound.

[mcp5500]

Hello Opimist. I purchased ver 3 and have this issue. Working as a hardware engineer and working with softare development for may years. this is what I see. I am not trolling! I have purchased Ver1, 2 and now 3 and a Prepar3d supporter. The only reason I have AMD is because I did many a design with their products. So I decided to take a hit in P3D to stay loyal to AMD. the only competitor to Intel in the CPU market. I see this as stated above but lets say your right and they used old drivers. That is poor workmanship. In engineering, software design, or QA the primary rule is "never assume".

[starsailorzed]

Same problems.

New user with AMD R9 295 GPU. Win 10.

Really haven't even been able to check out the software at all.

Guess it's back to Xplane and FSX...

[hawx2k]

Don't do that :). Once you get v3 up and running with the 14.12. drivers it's an absolute dream to fly with.

Yes rolling back to 14.12. is a bit of a hassle, and I do hope LM fixes it quickly, but in the interim that works fine.

[Optimist]

Hi mcp5500,

then please apologize for being too harsch. I was indeed not referring to your first posting which was a clear and understandable statement, but merely to the following posts which added nothing further but speculations in regards to LM actively not wanting to support AMD. I could not think of a reason why they should actively keep half the market from buying their products. :-)

I think LM stated earlier that they did test the newer drivers, but only on older hardware. Of course I do not know what´s correct, but if they had used latest drivers for all their testing rigs, such an issue could not go unnoticed during quality phase in my opion. Surely, nobody can test all possible configurations, but f.e. cards like the R9 200 series which has become quite a mainstream standard by now, should be included in the test sets, especially as they recommend 4GB+ cards . Maybe they were using some very specific funtionalities fo the 14.xxx drivers which are not supported anymore. Or maybe AMD really broke something which has only come up now with V3.

But as said, there is enough room for wild guesses, but in the end it is not helping and we will have to wait what solution LM comes up with. Important for me is only that they keep being transparent and at some point explain what the issue has been. And hopefully, that it will not happen again with each new graphics driver released... ^^

[pipe79]

ok i don't normally post at all , but i also am stumped how this has happened.

I also have the same issue but I'm running windows 7 , it loads in but within 10seconds it crashes , driver has crashed and recovered. i also have just brought into this and can't use it at all.

i have sent report to amd.

[mcp5500]

optimist.Let LM answer that. When I am told by LM to complain to AMD about this issue, when clearly that is their job, what other conclusion can I make?

[Bobcat999]

AMD R9 380 user here with Windows 7 64 bit. Same problem for me; P3D V3 crashes as soon as it starts to render the scenery whilst I am sitting on the runway.

First of all the graphics appear to render very slowly, then the mouse hangs so I cannot even access the menu, then the screen goes white for 15 seconds or so, then it goes black for 15 seconds or so, I press control alt delete to try to get Windows Task Manager open but the computer struggles as it is crippled by then. Finally after a long wait I get a box to say 'wait for P3D to respond or end process', and I have to end process of course. I think the graphics driver has reset by then. I must have been through this 20 or 30 times whilst tweaking things to try to get it to work - very fatiguing!

I just wish I had found these forums earlier, as I have wasted hours and hours and hours trying to solve this myself, and even swapped to an R9 280X that a friend gave me, so obviously I am not too happy about this, as it smacks of poor quality control.

As others have said, there is more than one graphics card manufacturer on the planet, and these R9 card are not exactly rare. To blame this solely on AMD is unfair.

I have never paid 199 dollars for software before, so I expected more. Whilst I understand there will always teething problems to be ironed out with new software, and I am sympathetic, this is a bit of a major one and should be deeply embarrassing for LM. With this show-stopper and all the potential issues getting add-ons to install even when it is fixed, I am now wondering whether it is all worth it.

Anyway, what I look for now is an excellent customer service response to win confidence back. I work in Aerospace also. If we create a problem for our customer we are expected to make a daily progress call to the customer to let them know what we are doing about it. As a minimum, I would expect a company of LM's stature to give us a daily progress report on here, as we have paid $199 for something we cannot currently use.

As for telling people to install 14.12 drivers, thanks for the info, but I can't as I need the newer drivers for other software I run!

Sorry for the long post, but I hope you can understand the frustration and the fatigue from being up until 3 in the morning trying to get this thing to work on my own, basically because I thought it was me or my computer at fault, as I honestly didn't think it could be something like this issue slipping through the net.

Come on LM - please help us out on this! Really tired now - off for a sleep!

[starsailorzed]

Same issue with Windows 10 and R9 290X.
Downgraded driver to 14.12 from here as a workaround:
http://support.amd.com/en-us/download/d ... &rev=14.12

ty for posting this link, Pober...I'm a technophile/computer "expert" and i STILL had a freakin hard time finding where to download this from without wasting oodles of time (which we should NOT be doing after buying such an "advanced" piece of software---if we were in beta and this were free, it'd be one thing, but this is a horrible experience thus far).

[starsailorzed]

AMD R9 380 user here with Windows 7 64 bit. Same problem for me; P3D V3 crashes as soon as it starts to render the scenery whilst I am sitting on the runway.

First of all the graphics appear to render very slowly, then the mouse hangs so I cannot even access the menu, then the screen goes white for 15 seconds or so, then it goes black for 15 seconds or so, I press control alt delete to try to get Windows Task Manager open but the computer struggles as it is crippled by then. Finally after a long wait I get a box to say 'wait for P3D to respond or end process', and I have to end process of course. I think the graphics driver has reset by then. I must have been through this 20 or 30 times whilst tweaking things to try to get it to work - very fatiguing!

I just wish I had found these forums earlier, as I have wasted hours and hours and hours trying to solve this myself, and even swapped to an R9 280X that a friend gave me, so obviously I am not too happy about this, as it smacks of poor quality control.

As others have said, there is more than one graphics card manufacturer on the planet, and these R9 card are not exactly rare. To blame this solely on AMD is unfair.

I have never paid 199 dollars for software before, so I expected more. Whilst I understand there will always teething problems to be ironed out with new software, and I am sympathetic, this is a bit of a major one and should be deeply embarrassing for LM. With this show-stopper and all the potential issues getting add-ons to install even when it is fixed, I am now wondering whether it is all worth it.

Anyway, what I look for now is an excellent customer service response to win confidence back. I work in Aerospace also. If we create a problem for our customer we are expected to make a daily progress call to the customer to let them know what we are doing about it. As a minimum, I would expect a company of LM's stature to give us a daily progress report on here, as we have paid $199 for something we cannot currently use.

As for telling people to install 14.12 drivers, thanks for the info, but I can't as I need the newer drivers for other software I run!

Sorry for the long post, but I hope you can understand the frustration and the fatigue from being up until 3 in the morning trying to get this thing to work on my own, basically because I thought it was me or my computer at fault, as I honestly didn't think it could be something like this issue slipping through the net.

Come on LM - please help us out on this! Really tired now - off for a sleep!

standing O, my man (or lady)...Standing O.

[Beau Hollis]

For version 3, we made a major overhaul to our GPU particle system to move more of the work off the CPU and onto the GPU. It appears from the content error reports posted that the AMD driver issue is related to the particle sorting that we now do on the GPU in a Compute Shader. The shader compiles without error, using the DirectX compiler, and executes in DirectX debug mode without throwing any validation errors or warning that might indicate that the pipeline state is improperly set. The error seems to be isolated to a specific family of hardware (eg R9) on a specific range of drivers (15.x). In cases like this, the issue if often in the driver where the generalized DirextX instructions are further compiled into hardware specific instructions. Often optimizations made to speed up one application will inadvertently cause some valid code from another application to be translated into invalid hardware instructions. Without access to the drivers source, there isn't much more we can do to track down the source. We have emailed our contacts at AMD. As soon as we receive any update from AMD regarding the cause of the bug or a timeline for a fix, we will post it here on our forums. Once we know what the driver bug is, we may be able to modify our shader to work around the issue, so that our users don't have to wait for a driver release to get up and running.

We did test with up-to-date drivers on a range of consumer and professional grade hardware from both AMD and NVidia, but do not currently have an R9 290X in our test lab. We apologize that this was not discovered sooner. We will specifically seek out high end AMD users in future betas to ensure this does not happen again.

[hawx2k]

Thanks Beau for the update. That's much appreciated. We are looking forward to a speedy resolution. My offer to help still stands. So if you want me to try and expedite things on AMD's side, shout.

Thanks,

Jan

[Saul]

If you're having this problem. Please can you take a little time and contact AMD reporting this issue. The more people contact AMD the more pressure they will be under to get a fix out.