How AMD Radeon Sabotages Itself & Its Partners: Development Timelines & Failures

[acroig]

"AMD's Radeon division keeps sabotaging itself and its partners, and to help illustrate how the VBIOS change does that, we've recapped the manufacturing process."

[Nunz]

I giggle when people say AMDs software is on par with NV. The Radeon VII launch proved otherwise. The 5700XT launch proved otherwise. And now the 5600XT launch has once again, proven otherwise.

The CPU division needs to show their GPU guys how to launch a product. The 5600XT is a badass card and should dominate that lower market segment, but AMD has bungled their last 4 GPU launches so hard it's self-sabotage.

[bill dennison]

Originally Posted by Nunz

I giggle when people say AMDs software is on par with NV. The Radeon VII launch proved otherwise. The 5700XT launch proved otherwise. And now the 5600XT launch has once again, proven otherwise.

The CPU division needs to show their GPU guys how to launch a product. The 5600XT is a badass card and should dominate that lower market segment, but AMD has bungled their last 4 GPU launches so hard it's self-sabotage.

what does a card launch have to do with driver problems

my gtx 680's had tons of problems at launch and took NV almost 6 months to fix all of them

and I seem to remember something called a GTX 970 3.5GB

they did better with my 980's but still had a few problems if I remember right

yes AMD stepped on their sword here they will get over it


………

and as for my 2080 ti

RTX has delayed more games than are out

and DLSS is a ****en joke at 4k

[Nunz]

Originally Posted by bill dennison

what does a card launch have to do with driver problems

Watch the video? Their last 4 GPU launches have been plagued by driver issues of all kinds; overclocking literally NOT WORKING AT ALL due to the Radeon software, games not launching or graphical anomalies (Eisberg is an example, and this is months after release..), CTDs, weird issues with Windows displaying odd colors during bootup or first install of drivers..

Then we get to the Radeon VII, a card that was essentially unusable by people who bought one on release day for weeks. Not once have I ever had an NV card in that kind of state and I've more than a fair share on release.

my gtx 680's had tons of problems at launch and took NV almost 6 months to fix all of them

I had 2 GTX680s on launch and rarely ran into issues that weren't correctable myself, and that was while running SLI. While I've certainly had inconveniences with NV drivers at times (and that hasn't been an issue since the last cards I ran SLI with, aka the GTX680s), I've never had something that I couldn't work around through NVInspector.

and I seem to remember something called a GTX 970 3.5GB

Hardware issue, not a driver issue. The card still functioned and worked, even though that was shady marketing and NV was justifiably crucified for it.

yes AMD stepped on their sword here they will get over it

Will they? Because they've made the exact same mistakes with 4 consecutive launches. Clearly it's not clicking in that office, or they just don't care about QC in the GPU division.

AMD needs to be majorly criticized because they've done an absolutely garbage job of QC and preparation for their GPU launches. Their CPU side has batted 1000 and knocked homeruns for 3 releases now - there is no more excuses, no more "Poor AMD.." crap. These guys are failing the company by botching their launches because they can't put out stable Day1 software. It's unacceptable.

I want to buy a Big Navi or an AMD card when they finally take the performance crown again (and they will, it's just a matter of time), but if their cards are releasing like this on launch, I'll pass. I haven't had a single driver issue in years, pretty much ever since I stopped running mGPU.

and as for my 2080 ti

RTX has delayed more games than are out

and DLSS is a ****en joke at 4k

This has nothing to do with the topic at hand, Bill, even though I agree with you.

[bill dennison]

4 consecutive launches maybe my last AMD card was the fury x

and my fury x cfx was fine driver wise as good as my 980 SLI
and the one fury x is still working fine in my sisters system

I think they have been cleaning up Raja messes with the last 4 launches

even with this coming big navi

after Raja left I think they started over almost from scratch so I don't look for a real high end AMD GPU without Raja's influence till next year

…………...
he was to have Intel's GPU out this year I don't see it happening

and then we will have to see if he used AMD GPU technologies to do it without steeping on patents

[pax]

I havent had any driver issues but I run the vega 64 and previous rx 480 on DP and I think most of the bugs were on hdmi...

Last few drivers have really amped up the bug patches tho.

[bill dennison]

Originally Posted by pax

I havent had any driver issues but I run the vega 64 and previous rx 480 on DP and I think most of the bugs were on hdmi...

Last few drivers have really amped up the bug patches tho.

I have always been on DP also

but thinking of going to 55 inch HDMI 2.1 soon as I can get a card with it
a AMD card as fast as my 2080 ti strix non RT or a 3080 ti
so I hope they work that out unless I can get a 55 inch with DP

I also don't overclock on the CPU at all and only very rarely on the GPU
I think a lot of people blame drivers for bad OC's

[Bitey]

LOL

[pax]

Ya ocing rates are close to the return rates of cards. They had one link for that that showed about 2% for amd and 1 % nvidia return rates... So few oc now I dont anymore cept on vcore undervolt that automatically gets me better boosts.

[Nunz]

Originally Posted by pax

Ya ocing rates are close to the return rates of cards. They had one link for that that showed about 2% for amd and 1 % nvidia return rates... So few oc now I dont anymore cept on vcore undervolt that automatically gets me better boosts.

Where are you getting these numbers from?

[pax]

https://www.reddit.com/r/Amd/comment...and_2080_from/

But varies from batch to batch...

[Nunz]

Thanks. What about the "OCing rates"? I'm just curious

[NWR_Midnight]

I have been holding off commenting on this. Partly because most problems people have with drivers, games, etc that they blame on drivers is actually caused by other software, the user, or windows itself. Even those that AMD acknowledges and takes responsibility for.



If you take a look at Nvidia's driver details, they have a fixed list (AMD has this), a known issue list (AMD has this), a new additions/options explanation detail area (AMD has similar when new features or changes come out), and one thing that AMD doesn't have, and I don't think ever will. A section that BLAMES issues on Windows/Microsoft.



Now, my son has an Nvidia card in his machine. He has issues once and a while. Usually resolved by installing the latest drivers, or a game update (but not always).

I have AMD, I generally don't have any issues. When I do, it is usually caused by me, because I am always tinkering with stuff. Or, it's a bad game update. There have been the rare occasions that it was a driver issue. Very rare.



This leads me to this past couple weeks. I have had a couple issues with a few games. Mainly really bad stuttering in Rainbow Six Siege for months now after one of their game updates, so I was always blaming the update, sometimes every few seconds, sometimes every minute or so.

I spent the last 2 weeks trying everything to track it down, every fix on the internet (something I have tried before in the past) I was to the point of blaming the AMD/Drivers for a change. (the main reason I hadn't responded to this thread).



Well today, I finally said screw it, I am going to completely wipe my machine and do a re-install. Something I really didn't want to do. But I did. I first tested the games that gave me problems after windows was done installing and updated. No issues, ran perfect. I installed chipset drivers, tested again.. NO issues. I installed the GPU drivers, No issues. So, Drivers have been eliminated.

list of order of install and testing result after every new installation of software:

Chrome - no issues
Aida 64 - no issues.
Thunderbird - no issues


Microsoft Office 2016 and Adobe reader (had the key for office saved to pdf as it was a digital purchase) - all issues appeared -read below

So, I was ready to blame Office or Adobe. I restored back to before Aida 64 was installed (forgot to save a point before Office install). Tested to make sure all worked properly - It did. I was about to try to install office 16 again, but realized that I did one step and didn't test before the install. Because my installation files for office 2016 are stored on my home server, I had to turn on file sharing to access the local network drives. I did not test any game after turning on file sharing, prior to installing Office.

This time I turned on file sharing and did my tests, and found out that it is either Microsoft's file sharing or something related to it. Still working on how to fix this, and a little confused, because all is fine until I turn it on after restoring back to prior to turning it on. But after turning it on, the problems start, and continue even if I disable it, but not restore back to prior.

I would have never found the right path of the cause if I didn't go step by step with testing at every step. I almost missed it because I didn't test after the simple step of turning on file sharing.



The moral of all of this is, don't be so quick to point fingers at Drivers. Even if AMD takes the blame. It can be as simple as a common setting in windows causing it.

[bill dennison]

Originally Posted by NWR_Midnight

I have been holding off commenting on this. Partly because most problems people have with drivers, games, etc that they blame on drivers is actually caused by other software, the user, or windows itself. Even those that AMD acknowledges and takes responsibility for.



If you take a look at Nvidia's driver details, they have a fixed list (AMD has this), a known issue list (AMD has this), a new additions/options explanation detail area (AMD has similar when new features or changes come out), and one thing that AMD doesn't have, and I don't think ever will. A section that BLAMES issues on Windows/Microsoft.



Now, my son has an Nvidia card in his machine. He has issues once and a while. Usually resolved by installing the latest drivers, or a game update (but not always).

I have AMD, I generally don't have any issues. When I do, it is usually caused by me, because I am always tinkering with stuff. Or, it's a bad game update. There have been the rare occasions that it was a driver issue. Very rare.



This leads me to this past couple weeks. I have had a couple issues with a few games. Mainly really bad stuttering in Rainbow Six Siege for months now after one of their game updates, so I was always blaming the update, sometimes every few seconds, sometimes every minute or so.

I spent the last 2 weeks trying everything to track it down, every fix on the internet (something I have tried before in the past) I was to the point of blaming the AMD/Drivers for a change. (the main reason I hadn't responded to this thread).



Well today, I finally said screw it, I am going to completely wipe my machine and do a re-install. Something I really didn't want to do. But I did. I first tested the games that gave me problems after windows was done installing and updated. No issues, ran perfect. I installed chipset drivers, tested again.. NO issues. I installed the GPU drivers, No issues. So, Drivers have been eliminated.

list of order of install and testing result after every new installation of software:

Chrome - no issues
Aida 64 - no issues.
Thunderbird - no issues


Microsoft Office 2016 and Adobe reader (had the key for office saved to pdf as it was a digital purchase) - all issues appeared -read below

So, I was ready to blame Office or Adobe. I restored back to before Aida 64 was installed (forgot to save a point before Office install). Tested to make sure all worked properly - It did. I was about to try to install office 16 again, but realized that I did one step and didn't test before the install. Because my installation files for office 2016 are stored on my home server, I had to turn on file sharing to access the local network drives. I did not test any game after turning on file sharing, prior to installing Office.

This time I turned on file sharing and did my tests, and found out that it is either Microsoft's file sharing or something related to it. Still working on how to fix this, and a little confused, because all is fine until I turn it on after restoring back to prior to turning it on. But after turning it on, the problems start, and continue even if I disable it, but not restore back to prior.

I would have never found the right path of the cause if I didn't go step by step with testing at every step. I almost missed it because I didn't test after the simple step of turning on file sharing.



The moral of all of this is, don't be so quick to point fingers at Drivers. Even if AMD takes the blame. It can be as simple as a common setting in windows causing it.

I had something like this with apple quicktime for windows and did the same thing to find it

with quicktime crossfire stopped working unload quicktime and it worked fine
have not used quicktime since

if you don't go step by step after a nuke and pave you can't just say " drivers "

[pax]

Originally Posted by Nunz

Thanks. What about the "OCing rates"? I'm just curious

A mention long ago on gamers nexus... "less than 5%"... very few oc.

[NWR_Midnight]

Originally Posted by bill dennison

if you don't go step by step after a nuke and pave you can't just say " drivers "

Yep! It took me most of the day, but I finally tracked down what was tied to network discovery/file sharing that was causing me my problems. It was the windows service SSDP Discovery. Once I disabled that, all was good.