What’s with all the nonsense drama from title commenters ? Those cards will still be supported thru a separate driver.
Having separate driver branches for different architectures is actually a great thing since they don’t have to worry that changes in the codebase for one architecture could introduce regressions or bugs in other one for example.
This should actually improve stability of both driver branches and benefit everybody.
If anything, it’s massively preferable to what they tried to do with the R9 Fury cards back in the day. Having a stable driver to lean on and then doing security updates is MASSIVELY preferable to releasing a driver package that’s completely unified that just so happens to be broken on a small set of GPUs.
I mention the Fury cards because anyone who owned them and used them on Windows throughout 2020 will tell you that the drivers were broken for them for a good 8 months. I bet money AMD does not want a repeat of that, so might as well set a stable driver in stone and give it security updates from there.
Well depends on how you look at it. When trying to prevent regressions or defects that affect older architecture hardware, what you say makes sense.
However in practice, spinning the driver off the mainline will lead to less attention to the spinoff. Much less attention when issues are raised, much less resources.
What’s with all the nonsense drama from title commenters ? Those cards will still be supported thru a separate driver.
Having separate driver branches for different architectures is actually a great thing since they don’t have to worry that changes in the codebase for one architecture could introduce regressions or bugs in other one for example.
This should actually improve stability of both driver branches and benefit everybody.
If anything, it’s massively preferable to what they tried to do with the R9 Fury cards back in the day. Having a stable driver to lean on and then doing security updates is MASSIVELY preferable to releasing a driver package that’s completely unified that just so happens to be broken on a small set of GPUs.
I mention the Fury cards because anyone who owned them and used them on Windows throughout 2020 will tell you that the drivers were broken for them for a good 8 months. I bet money AMD does not want a repeat of that, so might as well set a stable driver in stone and give it security updates from there.
Well depends on how you look at it. When trying to prevent regressions or defects that affect older architecture hardware, what you say makes sense.
However in practice, spinning the driver off the mainline will lead to less attention to the spinoff. Much less attention when issues are raised, much less resources.