Open Source AMD Driver Now Supports OpenGL 3.3 — and It's Getting Faster 100
An anonymous reader writes "With the latest open source Linux code published today the AMD RadeonSI Gallium3D driver supports OpenGL 3.3 and GLSL 1.50; this is the open source Linux graphics driver used for Radeon HD 7000 series and newer, including the new Hawaii GPUs. The OpenGL 3.3 support appeared in patches spread across Mesa and LLVM that should appear in their next releases. It was also found that the RadeonSI driver is becoming a lot faster and starting to compete with Catalyst, AMD's notorious Linux binary driver."
Re: (Score:1)
If that was humor it was a fail. Or maybe you were serious and don't understand what 'OpenGL version' means as opposed to other software ;)
Re: (Score:1)
At least Slackware is at 14.1. .. after a sprint through 4, 5, 6 to 7 to finally catch up to where Redhat was! ;D
Big version numbers - It matters!
Re: (Score:1)
Well, that and Firefox finally having catched up with the development speed of Chrome.
Firefox 26.0! ;D
Firefox - Browser for nerds, version numbering for idiots!
Re: (Score:2)
Bah, that's nothing - Less [wikipedia.org] is up to version 458. Get it together, Firefox!
Re: (Score:1)
Three metric Torvalds should be enough for anybody.
Re:Non-free Nvidia driver already at 4.4 (Score:4, Insightful)
How is it embarrassing? Several current and former AMD employees work on the Gallium 3D driver implementation as a side project, people like Tom Stellard. If anything, AMD is reaping the benefits of having opened up their hardware documentation.
Re:Non-free Nvidia driver already at 4.4 (Score:5, Insightful)
Non-free AMD driver is also up there somewhere. Can't find exact version for Linux but whatever, it's probably at 4.2 or later.
The problem is more that MESA only supports 3.3 - But the free drivers (e.g. Nouveau) does NOT support 3.3 so AMD is actually better at the moment. I do believe Nouveau will get 3.3 support soon however.
The real news here, though, is that performance of the free drivers are catching up to the proprietary drivers. That means AMD can ditch the proprietary drivers completely within a couple of years - which, if they can stay afloat that long, means great news for us Linux desktopers! :)
Re: (Score:3)
Catching up to the closed AMD drivers is a pretty low bar if we are talking about Linux performance...
Re: (Score:2)
Catching up to the closed AMD drivers is a pretty low bar if we are talking about Linux performance.
Depends on what we're talking. The first is simply catching up to the current standards - does the open source drivers even run the same code as the closed source ones. On this they're 3-4 years behind the "state of the art", OpenGL 3.3 was released in March 2010. The second part is catching up to Catalyst in performance - the open source team has said they don't have the manpower to make as many special cases as the Catalyst team, they're aiming at 60-70% performance. Maybe the open source drivers integrat
Re: (Score:2)
ATI/AMD promised decent performing open source drivers years ago. I would like to have seen this promise fulfilled much sooner. This is progress, but still not all they promised. Couple of years, sigh.
So I've bought in each camp. My 3 most recent computers--2 of which aren't very recent any more-- are an Intel CPU with Nvidia (a fanless GT 610, upgraded from the original 8500GT when the fan went bad), an AMD CPU with Radeon (HD 5400), and the newest is an Intel CPU with Intel's much improved HD line o
Re: (Score:1)
Then you want a Haswell chip with the HD4600 - any of the i5/7/E3 xeon's will do as the performance is much better then the HD4000. The only caveat to using a new Intel CPU is 8GB of memory is the minimum needed to really get any advantage from them. Otherwise stick with what you have
Re: (Score:2)
Nouveau can't even support a 1600x1200x32bit 75Hz refresh monitor. It puked for me on anything more than 1280x1024.
Absolutely useless.
Re: (Score:1)
Absolutely useless for me
FTFY
Re: (Score:2)
Re: (Score:2)
In arch linux it's as simple as
pacman -S nvidia
How did you manage to brick the PC btw?
Did it destroy your bios?
Re: (Score:2)
Re: (Score:2)
You can install the Catalyst drivers from RPMFusion non-free
yum install akmod-catalyst
No need to run horrible install scripts from self-extracting archives and bork your system.
Re: (Score:2)
Re: Non-free Nvidia driver already at 4.4 (Score:1)
Re: (Score:2)
I have found the nVidia drivers to be extremely easy to install. I have been running nVidia for years on several computers and generally their drivers just work. I have run into frequent problems with the Intel drivers such that on my netbook if power management kicks in on the screen it will never recover. Years ago I had a nightmare with the AMD drivers and Intel drivers. We had a computer where it was impossible to get the Intel drivers to work at all. We ended up buying a cheap nVidia card because their
Re: (Score:2)
Re: (Score:2)
The AMD installer runs under X. I had a previous AMD video card in here (from about 3 years ago), so I didn't have to do much more than install an updated driver. It doesn't look
4.0? (Score:1)
Re: 4.0? (Score:3, Informative)
Re: (Score:1)
If by "free" you mean "a lot more work" then yes. They can directly go to 4.
Re: (Score:1)
Getting everything done is the enemy of getting something done.
Getting 3.3 supported and then 4.0 is probably faster than going for 4.0 directly.
I know that it sounds retarded but in my experience trying to do everything at once usually puts you in a situation where you do the wrong things.
Re: (Score:3)
Getting everything done is the enemy of getting something done.
This works both ways, unfortunately.
Re: (Score:3)
To get to 4.0 they have to do all the work for the things in 3.3 AND all the work for things in 4.0 but not in 3.3. Therefore it makes sense to do all the work for the 3.3 things first so the driver can say "I support OpenGL 3.3" then after that, start working on the rest of the 4.0 support.
Re: (Score:3)
Re: (Score:2)
Re: (Score:1)
Why not trying having the tiniest clue on what the hell you're talking about?
Re: (Score:2)
Because the open source implementations are tied to the MESA project - whatever MESA supports is possible to support, although not required to support.
Anything MESA doesn't support, well, is impossible at the moment.
Re: (Score:2)
Hi Mikey!!
Re: (Score:2)
This is what I hate about GPU (opensource) drivers. Never EVER can anyone give full explanations on what the heck is going on.
Instead we get oblique hints which more or less equals "RTFS". Or in some rare cases, RTFM.
Every time I try to google this stuff up, I ragequit in despair after two hours.
Re:The firmware remains proprietary (Score:4, Interesting)
Releasing the firmware source would be pointless, since there is no available compiler which could target the hardware. They'd have to release the compiler and hardware specs as well.
Re: (Score:2)
And this is a problem...how, exactly?
Re: (Score:1)
Re: (Score:2)
And yet there was a time then IBM provided full schematics and BIOS listings to their original PC for the modest cost of the paper they were printed on.
Re: (Score:2)
It exponentially increases the amount of code that would have to be reviewed for proprietary secrets and patent infringement. On the other hand, just releasing the drivers wouldn't be as much of an issue, since they just target an interface that doesn't reveal what happens on the other side.
Re: (Score:1)
if you need to know something, ask in the proper mailling list... all this change everyday, mesa and graphic card drivers are under heavy developement, so the developers are the ones to know best what is going on (and even then, different developers work on different things)
Re:The firmware remains proprietary (Score:4, Informative)
This is what I hate about GPU (opensource) drivers. Never EVER can anyone give full explanations on what the heck is going on. Instead we get oblique hints which more or less equals "RTFS". Or in some rare cases, RTFM. Every time I try to google this stuff up, I ragequit in despair after two hours.
So you're saying you can't write Java without understanding how the JVM is built? The firmware provides you with a very low level API that is very similar to assembler, it's more like runtime-loaded microcode than normal code. If you really care to try, I suggest you start here [x.org]. Basically you place commands into a ring buffer that is read by the command processor (CP) on the graphics card and then executed on the GPU. There's a ton of registers you can set up, tons of commands, tons of formats (like all the texture formats) and while it is documented it's literally thousands of pages all together.
For example, for the Southern Islands generation alone there is:
229 pages of 3D register documentation
298 pages of instruction set architecture
49 pages of programming guide which expands the
54 pages of evergreen/cayman programming guide which expands the
43 pages of R600/700 programming guide.
Those 700 pages only walk you through the very basics of programming the GPU though, like assembler for a CPU. Beyond that there's very little in the way of tutorials, look at the existing source and figure out what it does down to the registers it sets and commands it sends. By the way, if things are not done in the right order the behavior is often undefined and may lead to soft or hard lock-ups. Personally I gave up because I realized the massive complexity of a modern GPU, quite frankly programming it at this level is extremely difficult.
Re:The firmware remains proprietary (Score:4, Insightful)
Both hardware and firmware are proprietary, the main feature of an open source driver is that it replaces binary blobs in kernel space. Basically it makes it easier for kernel developers to track down bugs since they can debug everything that runs within kernel space. While the buggy firmware can still kill your GPU that is isolated from the remaining system and easier to track down than a driver with write access to everything.
competing with catalyst (Score:1)
Fuck a duck they are setting the bar pretty low if their goal is to compete with AMDs Catalyst Drivers, the official AMD Drivers on both Windows & Linux are junk.
AMD should be ashamed of their official drivers, you would think after all these years they would have got a grip of their developers & started to produce some decent reliable drivers but alas this is not the case you have to manually delete files when removing the drivers as the drivers are too fucking dumb to remember where they installed
Re: (Score:1)
That's what I was going to say.
"It's easy when you aim low. How are your drivers doing against Nvidias Linux drivers AMD?", though my impression is that they have improved (also relative.)
I have no experience actually using them so I don't know how much suffering it is (even with a good distribution with a drivers package?) and I don't know how much blame should be put on Linux & others (You're of course free to argue that it wouldn't be a problem if the drivers was open-source.)
Re: (Score:2)
AMD Catalyst drivers are garbage. The other day I installed the latest version (13.something?) on my Samsung notebook and then decided to uninstal them and go back to the old factory video driver version. What happened? Firefox stopped working with some strange errors, IE, and a lot of other applications. It's unbelievable that nearly two decades after Windows 95, there still exist device drivers that can destroy your OS install and force you to install everything. You hear me AMD developers? Your drivers a
Re: (Score:1)
Buck a Fuck if you think the AMD driver is shit on Windows - it's all the Ducky Poo Catalyst Control Center that quacks like a moose. I never install anything more then the Radeon Driver from Windows update as it doesn't add the moosely Catalyst Control Center to the system and the latest version depends on dotnet 4 just to work on a windows box.
From a user standpoint, I've found that the Open Driver works well enough for me to get shit done and even game a bit under Wine - framerates can and do drop below
Re: (Score:1)
An interesting strategy indeed.
I hereby present my constant-time prime factorization algorithm. It works by always returning 2 and 3. Sure, it's not correct (unless you've tried to factorize 6), but it's fast as hell.
Re: (Score:2)
I'm amused that this is still even an issue.
AFAICT, the state of the art of open Linux video drivers hasn't actually advanced, in the relative scheme of things in at least fifteen years: Things still just barely work, doing somewhat new things, at best.
(Oh, sure: The desktop can be stable...sometimes. But I had a stable...sometimes desktop in 1999, too.)
Oh, I see it's exactly the opposite. Sure, the graphics drivers are lacking some features and performance, but there's a lot of energy in development. The Freedesktop and Mesa guys react friendly to bug reports and things get fixed. Also Mesa has paid developers from VMware, Intel and Red Hat working on the project. Also Wayland is advancing quite nicely. On the other hand, if we look at the Linux desktop environments, they are the areas where I hands down see the most bugs. KDE seems to have the best quali
Re: (Score:1)
I've been impressed with Fluxbox both the Dev (Yes single dev) and the stability. Sure it's not a full blown destop environment like KDE/Gnome but you know what? I don't need a full bloated DE either. What I want/need is the god damn window manager to get out of my way while offering multiple desktops so I can configure my work flow as needed on each one and fluxbox does all of this for me.
Very few flaws and when I actually reported a bug to the dev, pretty quick turn around on figuring out what the cause w
Re: (Score:2)
Re: (Score:1)
'Stable sometimes' is remarkable in the absence of hardware MFGs support. It hasn't really been that long since hardware MFGs have gotten behind support for open source drivers. Now that a AAA game developer / publisher / marketplace has come to the free-side of the source, things are looking bright indeed. [pcworld.com]
Folks are waking up to the reality that it might not be in the best interest of their business strategy rely too heavily on another company for their success. Especially not after MS's reception of W8,
OUYA fizzled (Score:2)
Starting a new software project (esp. game) it costs nothing extra to ensure it runs on all the major platforms.
I don't see how that's the case. It costs a lot of overhead money to target the major consoles. Console makers have been more interested in poaching established studios from other platforms than in nurturing startups. And until very recently, an indie studio had to lease a dedicated office rather than operating out of the developers' homes. OUYA is more open than the other consoles of its generation, with an API familiar to Android app developers and official sideloading, but it appears to have fizzled. Wha
Re: (Score:2)
I got modded -1 on my original posting, apparently because the yungin's here are too proud to remember their heritage.
Maybe you remember things differently than I do, but back in MY day hardware either came with (printed!) documentation, or had such documentation a letter away.
Now, granted: We didn't have massively-parallel graphics
Better than monitor rate. (Score:4, Informative)
Looking at those graphs, for those games, the current open source driver is running above the refresh rate of most monitors.
So while the catalyst driver may be faster, in some cases doubling the frame rate, I highly doubt you'd actually notice the difference.
Re: (Score:3)
Looking at those graphs, for those games, the current open source driver is running above the refresh rate of most monitors.
So while the catalyst driver may be faster, in some cases doubling the frame rate, I highly doubt you'd actually notice the difference.
Incorrect; it eliminates the need to run with vsync disabled and the subsequent screen tearing that causes...
Re: (Score:1)
The only time V-Sync needs to be used is for an SVGA (analog) connection. Otherwise it slows down the redraw of a Digital Connection such as DVI/Display Port/HDMI. So anyone with one of those connections to their display using V-Sync is an idiot/noob that needs to be gutted.
I've been on a DVI connection for over 3 years and disabled v-sync as soon as I moved. Never seen any tearing even when the game drops below 20FPS (I've seen as bad as 5FPS) but I know my card is obsolete (Radeon 5670 w/512).
Re: (Score:3)
Re: (Score:1)
The only time V-Sync needs to be used is for an SVGA (analog) connection. Otherwise it slows down the redraw of a Digital Connection such as DVI/Display Port/HDMI. So anyone with one of those connections to their display using V-Sync is an idiot/noob that needs to be gutted.
I've been on a DVI connection for over 3 years and disabled v-sync as soon as I moved. Never seen any tearing even when the game drops below 20FPS (I've seen as bad as 5FPS) but I know my card is obsolete (Radeon 5670 w/512).
VSync is for when games are rendering faster than the monitor not slower. You will see tearing if you have a game running at 90fps on a 60Hz monitor.
Re: (Score:2)
Re: (Score:2)
No, the frame rate doesn't matter. Sure you don't want to drop below 60fps. But that doesn't mean you need to render more than 60fps.
Above 60fps, without waiting for vsync, only a fragment of your rendered frame can possibly be displayed. I have wondered if graphics drivers can take advantage of this to only run the final fragment shader passes for the fragment of the screen that will be drawn.
Re: (Score:2)
Above monitor performance FPS seems useless until you factor in multi-monitor, screen resolution and multi-boxing. Or that games are more than movies (looking at you Japanese RPGs) and have to actually take input and do processing in between frames. Being able to drop a few frames for better input might just mean that click that keeps you alive makes it into the game. A
Re: (Score:2)
Input events are usually time-stamped by the operating system interrupts that captured them. On network games, these timestamps can be used by the server to work out after the fact who shot first. While pro-gamers like to say that frame rates higher than the monitor refresh rate help them win. I have yet to see a double blind study to confirm that it actually helps.
Usually network latency is a much bigger and more noticeable problem for this type of after the fact simulation. But the process can be applied
Re: (Score:2)
Of course, since the only games on the list are Quake 3 and Doom 3, games from 1999 and 2004 respectively.
A modern graphics card can finally run 15-year old games at max speed! What an impressive feat.
Radeon driver is working pretty well now (Score:2, Informative)
I was surpriced how well radeon driver is working. I'm not even using the newest version, but still the driver works considerably better on opengl use cases than it did just few years ago.