Slashdot Mirror
XGI, VIA Release Open Source Drivers
An anonymous reader writes "XGI has announced the release of open source drivers for its Volari family of graphics adapters. Efforts at X.Org to merge the new code into the head branch are already underway. Almost simultaneously, VIA has announced the immediate release of open source drivers for S3 Graphics UniChrome, VIA ProSavage and ProSavage DDR. Could these moves signal the beginning of a period of rapid improvement in Free drivers for video cards?"
So based on this news what is the best card to buy?
It isn't like they have much to lose. They sell hardware, the drivers simply let that hardware operate. They probably also know this will earn them points with the Open Source Community.. which is always a good thing.
Proceed with Format (Y/N)? Y
I wish all other hardware companies did this.
This is good news for fanless C3-based systems using CLE266 MPEG acceleration hardware: Via had released closed-source drivers (and, indeed, forked Xine to use them in a product called VeXP). These drivers were reverse-engineered to support an open source equivalent, which was less than completely reliable.
You could've hired me.
You paying attention to this ATI?
The quality of these drivers becomes extremely good in the X tree such that both ATI & Nvidia lose sales because of it.
The only way those two will release their own drivers as open-source is when they feel a pinch in the pocketbook.
...to the question about free drivers: yes, maybe, I certainly hope, all of the above.
We already get our drivers for $free on most platforms. Free as in open and easy, I certainly hope so. While we've made great strides in compatibility over the years, getting the hardware people on board and co-operating is still lagging.
When 95% of each sector of the hardware market is co-operating, then we'll just have to hope coders are doing something useful with the platforms now that they're working.
I'll put this in my personal "Good News" category for future reference.
If my grammar and spelling are off, I am [distracted/tired/careless] (take your pick)
"Could these moves signal the beginning of a period of rapid improvement in Free drivers for video cards?"
I doubt it. Just a coincidence. Wishful thinking. Once nVidia releases open source drivers, you may start to think otherwise.
Digital Sailor
If the performance on those cards is anywhere near decent I'll be buying one.
I don't know the meaning of the word 'don't' - J
I've heard that newer NVidia cards can boot straight to TV.
Now I just have to decide on whether I wait for someone to work out a Open source driver for the XGI card or just spend the ~$40 on a NVidia card when I have a perfectly decent XGI card already.
heh- who am I kidding. I'm cheap. And patient.
Come on guys- let's start reverse engineering these XGI drivers!
It was mentioned in a comment on slashdot somewhere that the code in the ATI or nVidia drivers may be propriatary and closed source as liscensed from someone else. They may have bought code from company x but company x may not allow for that code to be open source. So instead of re-writing the drivers entirely so that nVidia/ATI own all of their own code, they may just stick with the binary drivers to protect other companies IP.
Does it suck? yes, very much so, but the world is like that with software patents. Though I am not sure how those patents would apply in this case.
And guess what? I think this will ultimately increase their sales. As more drivers are available, choosing hardware to run Linux or some obscure OS won't be so difficult, so people will be more likely to buy a piece of hardware.
I think this will also improve the quality of their products. Often, drivers, like any other software, contain bugs, which can cause it to appear as if the hardware isn't working as well as it should. Or perhaps the driver isn't quite as efficient as it could be with system resources, so it seems as if the hardware isn't quite as fast as it should be. When these things are released under open source, it is more likely that things like this will get fixed and improved, and that will ultimately improve the vendor's hardware product without requiring any significant effort on the part of the vendor.
XGI and VIA are doing a smart thing. I'm heading over to write them an email about them and thank them. I suggest that others do the same. This is great news, and I hope other vendors will follow.
Could these moves signal the beginning of a period of rapid improvement in Free drivers for video cards?
Only if nVidia and/or ATI follow suit. (I know that in some cases they can't, but they could take an approach like Netscape and Sun did, release everything you do own and leave out the stuff you don't).
When looking through the kernel source code there is only support for 2D. Kernel bugreport X.org bugreport
Are we getting a fully open source GLX driver as well? Does that GLX driver exploit all the features of the chip? If there is hardware shaders can the GLX driver use those? Or are we getting something like the open source "nv" drive that only does 2D?
Unstable Apps: Our Android Apps Don't Suck
I'm glad to see this positive development from XGI. Releasing open source Linux drivers can't help but give them positive publicity. Back in November of 2003 I said we had "an opportunity to persuade [XGI] that supporting Linux by releasing drivers would gain them positive reviews and have an impact on sales." XGI has released the drivers, now it remains to be seen whether this drives sales.
I don't know who persuaded XGI to make this commitment to open source but I fully intend to consider XGI for my next video card. I'm using an nVidia card today on the basis of their closed-source driver support for Linux, but I'd rather support a company that embraces open source.
Read up on video cards and Macrovision ... if companies making popular cards like ATI and NVidia start open sourcing drivers, it'll allow end users to trivially circumvent DRM plans (also read up on the whole 'trusted path' idea ... encrypted straight from CPU to monitor).
It's time to shift your support to companies that support open source (and by consequence, oppose DRM).
The XGI drivers are 2D-only; you still have to use the binary library provided by XGI for 3D.
This sucks, too, because the performance of the XGI Volari V8 is comparable to a Radeon 9600 or Geforce 5700. And I'm sure that their drivers suck, so there's probably more performance in them. And it's dirt cheap, too. A 256MB card comes in at just under $100, and a 128MB card at $85.
XGI needs to be told that this isn't enough.
Why? Well, I have a bunch of machines here that use the KM 266 Via Chipset and support is scarce. Even proper 2D support is rare (I've only seen it in Fedora 3 and SUSE). Other than that, most other distros will either report 24 bit color (it's 32), or force you to use the generic VGA driver (which is SLOWwww).
And 3D support? Non-existant. Not that the 3D is spectacular on the KM series anyway, but it's certainly passable for screen savers, programs like Celestia, and other non-'Doom 3' purposs.
And it's not as though the KM 266 isn't capable of better. Under Windows it performs just great for what most of my users want - just not under Linux.
Thank you VIA, it will only help you...
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
...or you could look at it as giving people only what they are paying for.
Is it really a racket, then or is it actually more of a manufacturing strategy? On the surface, we all want to think that price should be based on what it costs to make it. But there's more to it and, really, the only time material, labor and overhead costs come in to play when pricing an item is finding out where your break-even point is. After that, it's essentially demand setting the price. The fact is, that there are several markets... the home user market, the professional user market and on and on.
And forget that you're a technical type and think like a business man who doesn't know tech. When you are told you have these three graphics cards to choose from, each with comparable capabilities, one of them has this ridiculously low price for its class. Are you inclined to buy that one? Most business people don't because it causes them to doubt the lower-cost device. "Why is it cheaper? Must not be as good."
But back to manufacturing, it's important to lower manufacturing costs where ever possible... if it were your job to do it, you'd probably be no different. It's cheaper to make a bunch of the same product and then disable features and sell them as lower-end rather than to manage that many more product manufacturing lines.
Is it frustrating to the technical consumer to know this? Hell yeah. I've got a Dell Inspiron advanced port replicator and a Dell Latitude advanced port replicator that are freaking identical hardware and they work interchangeably except that some ports don't function properly. I haven't decided to crack these two things open to find out what's different, but there is a fairly significant cost difference between the two devices.
Is it a racket? No... I think that goes a little too far.
The XGI release is 2d only (the kernel code is for fb support, not DRI), and from what I see on the Unichrome driver effort's mailing list archives the VIA source release is just making available to everyone what has been available through a "developer portal" for some time and does not make any more of the chipsets' features usable.
So the only possible real news here is a shift in the attitudes of these companies. We'll see how that works out in the future (whether enough information is released to allow open-source 3d drivers for XGI and full support for the VIA MPEG enc/dec acceleration).
who here has actually used an xgi card, and what was your experience?
I am extremely curious--they are cheap, but I want to know about performance.
"We don't know what we are doing, but we are doing it very carefully,..." Wherry, R.J. Personnel Psychology (1995)
I ask because outside nVidia, Windows drivers are usually lacking one way or another, specially when it comes to stuff like OpenGL. It could be interesting if someone worked independently on Win drivers for mainstream videocards.
PS: It's great to see some companies realizing they are on the hardware buissnes, not software. Thank you. I had interest in the S3 Deltachrome/Unichrome series, and now i'll most certainly try one out.
So, in return for them being one of the first gfx card makers to release OSS driver, possibly starting an OSS revolution in the graphics card/hardware industry, we melt their web server?
95% of all computer errors occur between chair and keyboard (TM)
They think their biggest asset is their feature complete and quite stable drivers, and that anyone can easily compete with them for hardware in the market. They feel their whole business model is based on "IP", which is to mean chip designs and driver source code.
In other words, they rely on a proprietary business model, and they wish to keep it that way, because it makes them more money.
Could these moves signal the beginning of a period of rapid improvement in Free drivers for video cards?
/. story here). for those who don't read the Developers section.
There's even open source hardware from the Open Source Project (OGP) coming out (info here and here, and the
The PCI version is due soon, and reported to have resolutions up to 2048x2048, dual-link DVI and TV-out (but won't be capable of playing HalfLife2 or anything like that).
SEO Copywriter. Just Say ON
Cool. XGI took a very old version of the SiS driver, stripped out many useful features, uglyfied the code beyond belief and calls this *their* effort of open source development.
And the worst part is that my name is all across "their" source.
Finally, probably needless to say, the 3D part is not included.
-- The Online Photo Editor - http://www.phixr.com
/*
* Copyright 1998-2005 VIA Technologies, Inc. All Rights Reserved.
* Copyright 2001-2005 S3 Graphics, Inc. All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sub license,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the
* next paragraph) shall be included in all copies or substantial portions
* of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
* VIA, S3 GRAPHICS, AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS IN THE SOFTWARE.
*/
I doubt this move will have any influence on ATI or Nvidia to open things up but we can always hope...
Well...I don't know about you, but I'm about to write ATI and nVidia letting them know that as soon as the open source drivers come out, I'm dumping my ATI and nVidia graphics cards, switching to hardware from people who fully support Open Source, and recommending to anyone who asks me that they do the same.
If enough people do that, perhaps ATI and nVidia will pay attention.
You may want to know that Matrox just released new Linux drivers for the first time in many months. Head over to their page.
This is not completely fair. Most hardware companies depend on code in their drivers that their staff did not write. As contracts generally go, the outside developer usually imposes limits on use and distribution of their work. It's invariably more expensive to purchase outsourced code without restrictions.
$5 / month hosted VPS on linux = awesome!
Won't compile with 2.6.11? Check.
Compiles with BIG, LARGE warnings about depreciated features being used in 2.6.10? Check.
Won't work under x64_64? Check.
2D part of drivers buggy? Check.
Infrequent releases that don't correct problems? Check.
No support for X RandR? Check.
Sorry, the ATI drivers don't pass muster. Perhaps I should've realized sooner with the constant weird 2D bugs I had with the ATI driver. Or the fact it wouldn't compile on 2.6.11. Or the fact it just plain won't work as advertised on 64-bit Linux.
I took out my Radeon 8500, put in a Geforce 2MX I had, and installed the nVidia driver. It was actually wrapped in an installer, rather than me having to manually untar and run scripts ala ATI. It asked if I wanted 32-bit compatibility OpenGL libraries. It told me that the 2.6.11 kernel fixed some AGP issues and was reccomended (which was good since I already had it, and only used the 2.6.10 because of ATI). X RandR started to work with the nVidia driver. 64-bit and 32-bit apps work flawlessly with each other.
ATI is shit. Their card hardware may be good, but without a driver, it might as well be an ISA SB16 for all the use I get out of it.
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
The current FOSS ATI X drivers are based on information (and, OTTOMH, code) released by ATI and good for cards up to the 9200 series. Higher-powered cards will run also, but the more advanced hardware features aren't used.
NVidia didn't even release the source to a commodity item like their nForce LAN chipset, so we had to clean-room our own for that one.
The Volari cards look good. I'm pleased that the hard-working lab-rats there have finally managed to convince management to Open their 3D drivers too (they Opened their 2D stuff more than a year ago). Now all I need to do is fine someone in Western Australia who sells the XGI cards other than as a novelty item.
The VIA S3 cards suck. S3 cards have always sucked, from their horrid little every-one-different pre-Virge series on down. The CyberBlades sucked less, but were still not in the same league as their competitors, not even on par with Intel's basic integrated chipset. At least now we might be better equipped to work around some of the suckiness.
Got time? Spend some of it coding or testing