Mesa Finally An OpenGL Implementation (On Intel Hardware)

Mesa 3D has famously always not been technically OpenGL (lacking certification), but times are changing: "This is a great day for Mesa and open-source graphics drivers. Just a tad over a month ago, I submitted OpenGL ES 2.0 conformance test results to Khronos for Intel Sandy Bridge and Ivy Bridge GPUs with Mesa 8.0.4. There were no objections during the 30 day review period, so we are now officially conformant! Finally being on that list is pretty cool. Not only is this great news for my team at Intel, but it's terrific news for Mesa. Mesa has had a long history with OpenGL, the ARB, and Khronos. This is, however, the first time that Mesa has ever, in any way, been listed as a conformant implementation. This is a big boost to Mesa's credibility."

Comments

Woohoo!

[Lord Lode]

But what about its speed?

What on NVidia and AMD cards?

Re:Woohoo!

[The Mighty Buzzard]

Everyone who plays 3D games or uses their gpu for any sort of distributed computing cares. Intel graphics solutions are simply not worth having on a primary desktop and I'm damned sure not going to stop doing the things I enjoy out of some fucked up sense of idealism. Most of us, even among linux users, aren't code-hippies. Sure, we'd prefer FOSS drivers but as long as there is a working solution we really don't care.

If you want to paint yourself a little sign and go picket outside their offices, have at it. The rest of us will be over here not giving a damn.

Great

[binarylarry]

It's awesome having a totally open source graphics stack, from Intel's open GPU drivers all the way up to stuff like the OpenGL interface.

The GPU packaged with the ivy bridge release is almost competent, so I'm excited to see what future generations will bring.

Re:Great

[hobarrera]

If Intel keeps up the pace, and nvidia/ati don't want to start loosing market, they'd better follow the example.
Intel is already taking over the medium-end market for non-gamers, and low-end market for gamers. Especially due to the huge power saving differences.

Intel support on MY Linux Box

[lord_rob the only on]

I had bought a motherboard with an onboard integrated Intel graphic chipset, and I've always had problems with it for everything concerning 3D (even a small transparent 2D window using the 3D chipset sometimes crashed).
Finally I've bought an ATI (the cheapest model as I don't want to play games) and I'm happy with it and free drivers.

What is Mesa?

[Hatta]

I never really understood what Mesa was. I thought it was what you installed if you wanted software rendering of OpenGL. If you wanted hardware rendering, you installed drivers for your hardware. But now Mesa is providing hardware accelerated OpenGL? What's the point if we have open source Intel drivers?

I don't get it.

Re:What is Mesa?

[binarylarry]

Mesa provides a software renderer. Technically, all opengl implementations are supposed to also provide software implementations of functions not supported in hardware. They're supposed to gracefully fallback to software (and the lovely performance that gets you).

Mesa's software renderer means that it's easier to write drivers, so you can have a fully working implementation as you reimplement various features with hardware support.

Re:What is Mesa?

[Bobakitoo]

I never really understood what Mesa was. I thought it was what you installed if you wanted software rendering of OpenGL. If you wanted hardware rendering, you installed drivers for your hardware. But now Mesa is providing hardware accelerated OpenGL? What's the point if we have open source Intel drivers?

I don't get it.

Mesa is the user-space API that talk to the driver. Some driver(eg: nvidia) may provide their own libGL.so, other may use libGL.so from mesa(eg: open source drivers). Also mesa can be use to render in software and even off-screen(eg: batch processing images files or video).

Re:What is Mesa?

[finity]

I've wondered this before too. How does it all fit together. I found this very helpful: http://blog.mecheye.net/2012/06/the-linux-graphics-stack/#rendering-stack

Re:What is Mesa?

[finity]

Here's the relevant part (sorry, editing on a phone isn't that easy...):

Now comes the fun part: modern hardware acceleration. I assume everybody already knows what OpenGL is. It’s not a library, there will never be one set of sources to alibGL.so. Each vendor is supposed to provide its ownlibGL.so. NVIDIA provides its own implementation of OpenGL and ships its ownlibGL.so, based on its implementations for Windows and OS X.

If you are running open-source drivers, yourlibGL.so implementation probably comes from Mesa. Mesa is many things, but one of the major things it provides that it is most famous for is its OpenGL implementation. It is an open-source implementation of the OpenGL API. Mesa itself has multiple backends for which it provides support. It has three CPU-based implementations: swrast (outdated and old, do not use it), softpipe (slow), llvmpipe (potentially fast). Mesa also has hardware-specific drivers. Intel supports Mesa and has built a number of drivers for their chipsets which are shipped inside Mesa. The radeon and nouveau drivers are also supported in Mesa, but are built on a different architecture: gallium.

Re:What is Mesa?

[MostAwesomeDude]

MesaGL is an implementation of the GL API that can use any of several backends to do its actual work, including a couple software renderers and also hardware renderers for many Intel, AMD/ATI, and nVidia chipsets. Your distribution probably splits each renderer into its own package for historical reasons.

Re:What is Mesa?

[Anonymous Coward]

> Each vendor is supposed to provide its own libGL.so
Wrong. Each vendor is supposed to provide a *_drv.so module which is loaded either by the X server (for indirect rendering) or by libGL (for direct rendering).
For indirect rendering, libGL doesn't need to know anything about the hardware the X server is using; it just converts GL calls to GLX protocol and sends it to the X server.
Unfortunately, nVidia just has to be different, so they provide their own libGL. At least newer versions can handle talking to X servers which aren't necessarily using nVidia hardware (an X client can talk to different X servers on different runs, or even talk to multiple X servers concurrently; what hardware the local X server is running only matters if you're talking to the local X server).

Re:Credibility?

[fuzzyfuzzyfungus]

In pure performance? Not particularly. In 'just supposed to work, and ships on some enormous percentage of all x86s churned out every year'? That's Intel.

Re:Credibility?

[JDG1980]

Not to be a jerk but.... Does anyone really think of Intel bestowing credibility in the graphics realm?

In terms of raw performance? No.

In terms of stability and compatibility? Yes. Keep in mind, Intel is the largest provider of x86 graphics hardware in the world.

Re:Credibility?

[Tough Love]

Intel graphics are fairly respectable these days. Still far behind AMD but not a joke any more.

What has /. come to

[marsu_k]

Now we have Jar-Jar Binks writing headlines

Re:Credibility?

[binarylarry]

That's because Linus is generally a pragmatist.

He wants shit to just work, he doesn't get religious about things.

Re:And why?

[TheCycoONE]

Mesa is the OpenGL state tracker for Gallium3D. Why throw out what you can still use?

Re:Credibility?

[deroby]

The older stuff, yes.
Those new core i3/i5/i7 CPU's : the integrated graphics have become VERY potent.

And yes, I have an i5 that comes with HD Graphics 3000 so I should know.

In fact, I also have an Nvidia NVS 4200M sitting in my Dell Lattitude that automatically kicks in when I require 3D stuff (read : games) and battery is not a concern. So for fun I started AION on the Intel hardware goofed around a bit, closed the application and then started it up again but on the Nvidia hardware (easy thanks to http://en.wikipedia.org/wiki/Nvidia_Optimus). Im not saying AION is the new standard to test things on, but for a 3D game it looks good (lots of details & eye-candy) and I simply happen to have it on my machine. (1680x1050 btw, I had the settings on automatic)
To be honest I'm a bit shocked to notice they both look VERY alike. The nvidia picture seems to have more hmm 'powerful colours' (?) while the intel gfx were a bit more washed out but then again seemed to have less visible edges (AA?) though some of the effects looked 'simplified'.
Otherwise, both maintained a steady 45+ fps which is more than good enough for me. wow.

If you're more of a numbers guy, feel free to compare yourself :
* http://community.futuremark.com/hardware/gpu/NVIDIA+NVS+4200M/
* http://community.futuremark.com/hardware/gpu/Intel+HD+Graphics+3000+Mobile

Seems they come pretty close to each other and the difference in DirectX version supported might explain the noticed difference in effects. They were still there, but just slightly less... hmm... complex.

Sure there are a lot more powerful dedicated gfx cards around (hey, laptop here!), but trust me when I say that for 95% of the market, these integrated graphics are more than sufficient ! If you're in the 5% of users that 'needs' a powerful 3D processor, then by all means do; but claiming all Intel IGP's are a piece of shit is like saying that you don't have a decent printer if you don't go for the Xerox iGen 150 !

Re:Why OpenGL ES?

[UnknownSoldier]

No. My day job is working with OpenGL ES 1.x, 2.x, and WebGL. You will want this link. Don't let the title through you -- WebGL *is* OpenGL ES 2.0
http://www.khronos.org/webgl/wiki/WebGL_and_OpenGL_Differences