AMD Will Release New Catalyst Linux Driver Update This Month

[RussianNeuroMancer]

Basically it was a 1 man project, and that man has left for a new job at HP.
DRI3 was always only half-finished, and no one wants to step up and finish it off - most of the devs who care about what it fixes are more focused on moving over to Wayland rather than trying to hack the same concepts into X.

Thanks for explanation!

[eydee]

the point is that people with AMD hardware haven't been able to use it.

I wonder how many people are there who have that one single card it supports, and also are using linux. Probably close to none.

[xeekei]

You're kidding, right ? Six months to write an entire open source driver stack from scratch with the same functionality as radeon / -ati / radeonsi / amdgcn (llvm) ?



I'm sure you understand that legal/IP review is required before going public, not just for going upstream but I feel the need to say it anyways

We are impatient. But take it as a compliment. Most of us really, really want to buy AMD. We don't like NVidia's attitude, but frankly, if you play heavy games, their stuff is just better right now. I'm sorry!

[nanonyme]

You're kidding, right ? Six months to write an entire open source driver stack from scratch with the same functionality as radeon / -ati / radeonsi / amdgcn (llvm) ?



I'm sure you understand that legal/IP review is required before going public, not just for going upstream but I feel the need to say it anyways

We do realize legal/IP review is an expensive and time-consuming process but as is there's no visibility or even tentative ETA's for the effort. That's not particularly nice

[bridgman]

We do realize legal/IP review is an expensive and time-consuming process but as is there's no visibility or even tentative ETA's for the effort. That's not particularly nice

Pretty sure I have provided tentative ETAs a couple of times already, but let me do it again.

We are trying to get the code out in time to make the 3.21 merge window if possible. Not sure if we will make it though.

[duby229]

Just nitpicking sorry,

Doesn't 3.21 = 4.1 as of right now?

[chris200x9]

Just nitpicking sorry,

Doesn't 3.21 = 4.1 as of right now?

I thought it was intentional, like a joke. "STOP ASKING ME FOR TIME FRAMES!"

[bridgman]

Just nitpicking sorry, Doesn't 3.21 = 4.1 as of right now?

D'oh !!

I thought it was intentional, like a joke. "STOP ASKING ME FOR TIME FRAMES!"

Nope, good idea though. Just a bit slow today

[profoundWHALE]

Is there anybody that knows how is the power usage of the open source drivers compared to catalyst? In particular while using the desktop with 3D effects enabled. Thanks!

This sort of measurement... doesn't work so well. You could have the open source graphics use more Watts, but not have as much CPU usage and having a lower latency (I say latency because I still get that weird jittering in every game with catalyst, regardless of settings). You could have catalyst pull more Watts but it turns out it had like 20% better performance. This is of course assuming that they're both running the same version of OpenGL. Maybe when running the open source graphics you use a legacy OpenGL which could give you excellent fps and lower power usage, but at the expense of how it looks or artifacts.

I'm fairly certain that the performance and features of the newer AMD cards will be mostly on par due to the whole new open source graphics driver that AMD is pushing.

As for the older cards... Keep your fingers crossed?

Just to be clear, it's not a "whole new open source graphics driver", it's a "substantially restructured open source kernel driver with new user/kernel interface plus additional code at the bottom of the existing open source userspace drivers to let them work with either old or new interfaces".

Well, when I clean my room of boxes and cables and clothes, it looks 'brand new' And it functions much better. It's A-okay with me if the new driver is just like that.

What I'm really liking is that it will be able to integrate so much better with the kernel, which should give better performance regardless.

It's certainly taking as much time as a "whole new open source graphics driver".

Over 6 months now and still nothing.

Well I also don't see an R9 390X so I'm in no rush. What, do you have the R9 285? I want to know why you care so much.

[Kano]

As i have got a Nvidia Optimus script ready to output Nvidia cards to Intel or AMD gfx ports (i know that there are no laptops with AMD+Nvidia, but it works). But why does fglrx not support this for the muxless systems? You only get an error in the xorg log file if you try to init fglrx on those, btw. some systems works with older fglrx drivers via hdmi, that does not work anymore... So only oss drivers can be used via prime.