Aron Schatz writes "VIA has announced that they will start a new site (http://linux.via.com.tw — doesn't exist yet) specifically for the development of open source drivers. From their press release: 'Over the following months, VIA will work with the community to enable 2D, 3D and video playback acceleration to ensure the best possible Open Source experience on VIA Processor Platforms. 'To further improve cooperation with the community, VIA will also adhere to a regular quarterly release schedule that is aligned with kernel changes and release of major Linux distributions. In addition, beta releases will be issued on the site as needed, and a bug report and tracking feature will also be integrated.' Nvidia should be next."
They've timed this to fairly well coincide with a new processor design [via.com.tw] that promises better performance than what they've had to date. Hopefully not just drivers but optimizations for their CPU will take off in maturity alongside the growth of their deployed footprint.
by Anonymous Coward
on Tuesday April 08 2008, @04:45PM (#23005828)
NVidia use a whole lot of IP from other vendors, and they CAN'T make everything open source because of that. NVidia has made massive efforts in helping their products work on linux as smoothly as possible, and should be praised for their efforts, not berated.
If Via own all the IP they use, great! They'll be able to open source the lot. NVidia doesn't, so can't, so why give them a hard time after all their efforts to open source as much as possible?
Because no one asked for their IP. All anyone wants is details on the API. They can keep their drivers.
Nvida has also not open sourced "as much as possible" They got dragged kicking and screaming every step of the way. They didn't assist with open sourced drivers for their on board chipset devices until the open source folks reverse engineered the NVidia's drivers and did a better job than NVidia did.
Intel can do it. ATI has promised to do it and now so does VIA. Why is NVidia different?
I'm about 4 months from my next video card purchase and I will be taking a hard look at who has the best Linux support. A 10 FPS difference is not worth drivers that seem to need reinstalling every reboot (thanks NVIDIA).
However, this new driver does support the Radeon HD 2000 (R600) family... Don't expect any miracles from this driver just yet. At this point, the RadeonHD driver is really targeted for developers and those wanting to use the experimental driver whether it is due to problems using the fglrx driver on the system or just wishing to test out the driver to see if it works for you. As long as AMD sticks to their word on delivering the rest of their documentation, there will not
Intel can do it. ATI has promised to do it and now so does VIA. Why is NVidia different?
Since September, ATI is really far along, to the point where the free drivers are already improving with the published specs. ATI isn't only promising, they are delivering.
by Anonymous Coward
on Tuesday April 08 2008, @06:39PM (#23006880)
> Intel can do it. ATI has promised to do it and now so does VIA. Why is NVidia different?
I'll probably get suicided for telling you this, but i cant stay quiet any more.
the driver would reveil that it only has two parameter: maxfps and minpowerreq. Those parameter, on the other hand, is set from a simple jumper on the GFX-Card - the GPU isself has been the same from the first GForce almost a decade ago to the newest 9900 Models they are going to "invent" next year. All GPU's have the same Core, reverse engineered from a UFO that crashed in Fairbanks, New Mexico. They would release the new Models now, but Sam Carter is in another galaxy and is only one who has a signature-key for the Asgarth compiler. So, sorry, no opensource drivers for you!...... what's that humming in front of my windows....?
A 10 FPS difference is not worth drivers that seem to need reinstalling every reboot (thanks NVIDIA).
Are you living on the bleeding edge of daily recompiled kernels or something? I checked the box in restricted-manager once, and I've never had to worry about it again. I suppose we'll see how it goes with the pending Hardy upgrade soon, but at least for close to six months now it's been a complete non-issue. I do wish the LCD TV I have on the second DVI port would work right though...
"A 10 FPS difference is not worth drivers that seem to need reinstalling every reboot (thanks NVIDIA)."
Let me guess, you use Ubuntu and have the restricted drivers installed?
I haven't looked at this for 1 or 2 years, but the restricted drivers (including the nvidia drivers) are copied to a ram disk mounted somewhere in the kernel module directory tree (can't remember where) at boot time (no idea why).
So, if you have the restricted drivers installed and then install the nvidia drivers from nvidia's web
It's likely that your reinstall on every boot is necessary because your distribution includes several open source replacement drivers which interfere with the one you installed. This can be fixed via adding
From what I can find online NOBODY has h.264 accelerated playback in Linux, and the best solution is to use a hack that loads a commercial software decoder made for Windows.
Until I read this I thought my next HTPC would be ATI if things ever came to be, but now it may be VIA.
I will build my next PC as soon as I find a fairly inexpensive card/chipset that will do HDMI out, and hardware H.264 in Linux. I don't care whether it is good closed source (like Nvidia graphics)or open source support either, Just that it works well and is fairly low power usage (pegging a fast multi-core CPU does not count).
Until Nvidia offers feature parity with Windows on Linux or enough is opened for the community to do it themselves they deserve a hard time.
If someone knows a good solution for this let me know so I can stop waiting.
We're seeing more and more VIA CPUs in Linux-based "low-end" laptops. I think this really bode well for Linux. If we establish a presence in these internet-as-an-appliance devices, we can use it as a staging point to move into the desktop market.
There's a reason (two, actually) 1: VIA chips, although they aren't known for stellar performance, are usually pretty low-power (but nowhere near ARM-based stuff) 2: VIA produced something called the Nanobook, a reference design for the subsub-notebooks. Two machines have been released under this design: the Everex Cloudbook and the Packard-Bell Easynote XS
You are right. The rise of the low cost computer is the beginning of the end for high profit margins on Windows. The only real way they have of combatting FOSS is to release Windows for next to nothing and hope to extract some sort of money with Office. But that is a losing tactic as well. As capable computers approach the cost of a VCR, consumers will have the choice of getting something that does the job, or something that does the job and costs twice as much.
What other options do MS have? They can't kill/buy off Via or the other commodity PC manufacturers that will spring up as know-how increases. Import restrictions can't stop world trade, are not popular and take a long time to implement.
Killer apps to increase future price of computers? The average human has only so many needs that can be filled by an increasing number of instructions per second, especially when those instructions must be executed in parallel. Computers are fast enough for our senses (HD video, sound) and communications needs (bottleneck is in the networking). Most of the killer apps are already here. Any extra functionality enabled by some sort of high powered Intel machine is a small percentage of total functionality provided by the current crop of computers.
In this sort of environment, owning a computer that will do almost everything is an easy decision for $200, even if it means acquainting yourself with a different operating system. Ubuntu is easy to install, but even easier if it just comes pre-installed on your device with all drivers working. With understanding comes trust, acceptance and consideration for use in other spheres of life. If anything, the killer app is a small, very low power, solid state computer (hence silent while requiring zero maintenance) - for firewalls, NAS, HTPC, portables, general PC use and home gaming. There is no real margin in any of this (except games), and no ability to fund the survival mechanisms of an operating system and office suite monopolist. Only niche players will find the margins.
With high profit margins comes the resources to lobby, to advertise, to muscle hardware vendors. Maintaining their monopoly has not been cheap, but while there was profit to be had it was a sound business proposition. Without the money coming in, the collapse will be reminiscent of the Soviet Union - without the resources to maintain the empire, the decay will accelerate rapidly and people will be surprised at how rapidly and pervasively it actually happens.
I think that would be terrific, but the announcement/article didn't say that... this is wild speculation as far as I can see.
ATI pissed me off too often where I had ATI in my Dell laptop for quite a while. I watched friends with nVidia in their machines do all sorts of nice things under Linux that I couldn't do with ATI. It was annoying, so eventually I bought an nVidia card for my laptop to replace the ATI and I too was doing nice things under Linux that I couldn't do with ATI... then ATI announced they were going strong on releasing their hardware specs and stuff like that opening the door for completely OSS drivers for ATI hardware. (I haven't seen anything yet, but I haven't been watching since I switched to nVidia.)
And now here I sit with nVidia hardware in my laptop... waiting for driver updates, features and bugs to be added... same-ole-same-ole. I'd love for the speculation that nVidia will essentially fold under the pressure, but at the moment, I don't see that they are showing any signs of pressure.
I've got an intel G35 (integrated graphics) based board from Asus (P5E-VM HDMI) and im'm prety happy with 3D at 1600x1200 on ubuntu 8.04beta.
I usualy don't like integrated devices, but I choose this board just bacause intel released open source drivers/specs and they deserve this choice.
then ATI announced they were going strong on releasing their hardware specs and stuff like that opening the door for completely OSS drivers for ATI hardware. (I haven't seen anything yet, but I haven't been watching since I switched to nVidia.)
Released: 2D specs for R300-R500, R600 3D specs for R300-R500 Not released: 3D specs for R600 TCore (graphics card simulation) Might be released: Low-end code from the fglrx driver
In short, they still haven't released the specs on their latest generation and R700 is expected sometime this year. Yes, it's a promising development but if you want the latest and greatest, it's closed source whether you go with ATI or nVidia...
While not exactly a major powerhouse by my estimate, VIA still holds a pretty decent hunk of marketshare. Nice to see one more (relatively) large player see Linux as a valid enough market to make this kind of effort.
I'll believe it when I see the drivers working on my Ubuntu system with desktop effects active. I've tried the OpenChrome drivers and other things and nothing works with the UniChrome Pro CN400/PM880 video card that I have. Via has been very disappointing so far.
I was also thinking this is too little, too late. After a few years of playing around with VIA systems, I've moved to Intel boards where opensource drivers just work.
Nevertheless, it would be nice to see this work for real. Competition is always welcome. Frankly, it's weird that VIA hardware is geared towards embedded/mobile use, while providing drivers mainly for the most power-hungry OS on the planet.
They go out of their way to encourage the growth of an ecosystem in which their products can compete. Not too long ago I hadn't heard of any PC/laptop processors besides AMD/Intel, but thanks to VIA's encouragement of the Ultra-mobile PC market (or 'netbooks' as Intel likes to call them) they have suddenly become a player.
VIA created the nanobook [via.com.tw] reference design for mini-laptops that use their low-cost, low-power chips. Already the CloudBook [wikipedia.org] has come out based on that design, and in other countries various similar laptops have been released from different distributors. Now they're stimulating essential linux development, which will continue to increase the value of their low-cost platform. This has "win" written all over it; we're all going to come out ahead thanks to their strategy.
I hope they consider extending it to their crypto accelerator. Even low end Via boards (like the C3 I bought two years ago for $60 from newegg) include a hardware RNG and low level AES routines, and it would be cool to get some proper support. I've used Sun's crypto accelerators on their T2000's and the difference on certain algorithms is stunning.
Umm, this has been in the vanilla Linux kernel for a while. I've tried it with the C7 with great success, too bad the rest of the motherboard wasn't particularly good with Linux.
by Anonymous Coward
on Tuesday April 08 2008, @05:17PM (#23006144)
The Via H/W random generator is used as a/dev/[u]random driver in current kernels, and the AES engine is available as an openssl 'engine', so I'm not sure what else you need.
Via really stuffed up, however, when they made repeated half-baked attempts at 'semi-binary' drivers which worked only on ancient versions like Redhat 9 and wouldn't provide any support or information on the MPEG decoding chip (in the CLE266 and above) which was essential to getting working DVD and DVB playback on the low power boards like the Nemiah.
Via really stuffed up, however, when they made repeated half-baked attempts at 'semi-binary' drivers which worked only on ancient versions like Redhat 9 and wouldn't provide any support or information on the MPEG decoding chip (in the CLE266 and above) which was essential to getting working DVD and DVB playback on the low power boards like the Nemiah.
I agree with the general sentiments on the VIA-Linux relationship. However, my old system with a 1 GHz Nehemiah [sic] and CLE266 played DVDs just fine without MPEG2 acceleration, using MPlayer.
Methinks "VIA will work with the community" translates to "VIA would really like the community to do all the work, and will be good enough to host it on their website", perhaps?
Methinks "VIA will work with the community" translates to "VIA would really like the community to do all the work,
If they just publish the specs and stand back, I'm fine with that.
If they also release some of their current code under a free license it's a nice bonus.... and will be good enough to host it on their website", perhaps?
Their press announcement arrived before any content, sigh.
The PR says the website will start with: drivers, technical documentation, source code, and information regarding the VIA CN700, CX700/M, CN896 and the new VIA VX800 chipsets. It'd be good to see docs on their more widely used chipsets, like vt8235 and vt8237... detailed ones, including errata. I mean, currently they piss off almost everyone who uses their chipsets, so why would anyone want to buy NEW hardware with VIA chips if it's not even clear the current stuff can be made to work well?
It's a nice idea, years overdue. But even at that, pre-announced.
Already big in little markets (Score:4, Informative)
Can you get it at ... (Score:5, Funny)
Can you get the Zonbu at zombo [zombo.com]?
Parent
Re: (Score:2)
Re: (Score:2)
Re:Can you get it at ... (Score:4, Funny)
Parent
Re:Can you get it at ... (Score:5, Funny)
Parent
Re:Already big in little markets (Score:4, Funny)
Parent
This is good timing. (Score:3, Interesting)
Nvidia have already open sourced what they can (Score:4, Interesting)
If Via own all the IP they use, great! They'll be able to open source the lot. NVidia doesn't, so can't, so why give them a hard time after all their efforts to open source as much as possible?
Re:Nvidia have already open sourced what they can (Score:5, Insightful)
Nvida has also not open sourced "as much as possible" They got dragged kicking and screaming every step of the way. They didn't assist with open sourced drivers for their on board chipset devices until the open source folks reverse engineered the NVidia's drivers and did a better job than NVidia did.
Intel can do it. ATI has promised to do it and now so does VIA. Why is NVidia different?
I'm about 4 months from my next video card purchase and I will be taking a hard look at who has the best Linux support. A 10 FPS difference is not worth drivers that seem to need reinstalling every reboot (thanks NVIDIA).
Parent
Re:Nvidia have already open sourced what they can (Score:5, Informative)
ATI hasn't just promised, they did:
http://ati.amd.com/developer/open_gpu_documentation.html [amd.com]
http://www.phoronix.com/scan.php?page=article&item=842&num=1 [phoronix.com]
Parent
Re: (Score:3, Informative)
Getting closer, but still no cigar:
Re:Nvidia have already open sourced what they can (Score:5, Informative)
Parent
Re: (Score:2)
The only time I reboot is when I recompile the kernel. Which will require rebuilding/installing the nvidia-drivers package anyway.
Re: (Score:2)
Re:Nvidia have already open sourced what they can (Score:5, Funny)
I'll probably get suicided for telling you this, but i cant stay quiet any more.
the driver would reveil that it only has two parameter: maxfps and minpowerreq. Those parameter, on the other hand, is set from a simple jumper on the GFX-Card - the GPU isself has been the same from the first GForce almost a decade ago to the newest 9900 Models they are going to "invent" next year. All GPU's have the same Core, reverse engineered from a UFO that crashed in Fairbanks, New Mexico. They would release the new Models now, but Sam Carter is in another galaxy and is only one who has a signature-key for the Asgarth compiler. So, sorry, no opensource drivers for you!
Parent
Re: (Score:3, Interesting)
A 10 FPS difference is not worth drivers that seem to need reinstalling every reboot (thanks NVIDIA).
Are you living on the bleeding edge of daily recompiled kernels or something? I checked the box in restricted-manager once, and I've never had to worry about it again. I suppose we'll see how it goes with the pending Hardy upgrade soon, but at least for close to six months now it's been a complete non-issue. I do wish the LCD TV I have on the second DVI port would work right though...
Re: (Score:3, Informative)
Re: (Score:3, Informative)
See:
http://wiki.debian.org/NvidiaGraphicsDrivers#head-6cb9442ef3215e7aa8e2e1a13c73a7819a9e9890
Re: (Score:3, Informative)
DISABLED_MODULES="nvidia nvidia_legacy nvidia_new"
to
Re:Nvidia have already open sourced what they can (Score:5, Interesting)
From what I can find online NOBODY has h.264 accelerated playback in Linux, and the best solution is to use a hack that loads a commercial software decoder made for Windows.
Until I read this I thought my next HTPC would be ATI if things ever came to be, but now it may be VIA.
I will build my next PC as soon as I find a fairly inexpensive card/chipset that will do HDMI out, and hardware H.264 in Linux. I don't care whether it is good closed source (like Nvidia graphics)or open source support either, Just that it works well and is fairly low power usage (pegging a fast multi-core CPU does not count).
Until Nvidia offers feature parity with Windows on Linux or enough is opened for the community to do it themselves they deserve a hard time.
If someone knows a good solution for this let me know so I can stop waiting.
Parent
Re: (Score:3, Insightful)
Re: (Score:2)
What are you smoking??
Since when DVI started supporting audio?
Year of the Linux.... portable? (Score:3, Informative)
Re:Year of the Linux.... portable? (Score:4, Informative)
It is the year of the rat:
http://en.wikipedia.org/wiki/Rat_(zodiac) [wikipedia.org]
Learn your lunar calendars, Slashdot.
Parent
Re:Year of the Linux.... portable? (Score:5, Funny)
Parent
Year of the rant (Score:3, Funny)
Re:Year of the rant (Score:5, Funny)
Parent
Re:Year of the Linux.... portable? (Score:4, Informative)
1: VIA chips, although they aren't known for stellar performance, are usually pretty low-power (but nowhere near ARM-based stuff)
2: VIA produced something called the Nanobook, a reference design for the subsub-notebooks. Two machines have been released under this design: the Everex Cloudbook and the Packard-Bell Easynote XS
Parent
Re:Year of the Linux.... portable? (Score:4, Interesting)
What other options do MS have? They can't kill/buy off Via or the other commodity PC manufacturers that will spring up as know-how increases. Import restrictions can't stop world trade, are not popular and take a long time to implement.
Killer apps to increase future price of computers? The average human has only so many needs that can be filled by an increasing number of instructions per second, especially when those instructions must be executed in parallel. Computers are fast enough for our senses (HD video, sound) and communications needs (bottleneck is in the networking). Most of the killer apps are already here. Any extra functionality enabled by some sort of high powered Intel machine is a small percentage of total functionality provided by the current crop of computers.
In this sort of environment, owning a computer that will do almost everything is an easy decision for $200, even if it means acquainting yourself with a different operating system. Ubuntu is easy to install, but even easier if it just comes pre-installed on your device with all drivers working. With understanding comes trust, acceptance and consideration for use in other spheres of life. If anything, the killer app is a small, very low power, solid state computer (hence silent while requiring zero maintenance) - for firewalls, NAS, HTPC, portables, general PC use and home gaming. There is no real margin in any of this (except games), and no ability to fund the survival mechanisms of an operating system and office suite monopolist. Only niche players will find the margins.
With high profit margins comes the resources to lobby, to advertise, to muscle hardware vendors. Maintaining their monopoly has not been cheap, but while there was profit to be had it was a sound business proposition. Without the money coming in, the collapse will be reminiscent of the Soviet Union - without the resources to maintain the empire, the decay will accelerate rapidly and people will be surprised at how rapidly and pervasively it actually happens.
Parent
nVidia next? (Score:3, Insightful)
ATI pissed me off too often where I had ATI in my Dell laptop for quite a while. I watched friends with nVidia in their machines do all sorts of nice things under Linux that I couldn't do with ATI. It was annoying, so eventually I bought an nVidia card for my laptop to replace the ATI and I too was doing nice things under Linux that I couldn't do with ATI... then ATI announced they were going strong on releasing their hardware specs and stuff like that opening the door for completely OSS drivers for ATI hardware. (I haven't seen anything yet, but I haven't been watching since I switched to nVidia.)
And now here I sit with nVidia hardware in my laptop... waiting for driver updates, features and bugs to be added... same-ole-same-ole. I'd love for the speculation that nVidia will essentially fold under the pressure, but at the moment, I don't see that they are showing any signs of pressure.
Re:nVidia next? (Score:4, Interesting)
Parent
Re:nVidia next? (Score:4, Informative)
2D specs for R300-R500, R600
3D specs for R300-R500
Not released:
3D specs for R600
TCore (graphics card simulation)
Might be released:
Low-end code from the fglrx driver
In short, they still haven't released the specs on their latest generation and R700 is expected sometime this year. Yes, it's a promising development but if you want the latest and greatest, it's closed source whether you go with ATI or nVidia...
Parent
kudos (Score:4, Insightful)
Re: (Score:3, Insightful)
I'll believe it when I see it working.... (Score:4, Insightful)
Re:I'll believe it when I see it working.... (Score:4, Insightful)
I was also thinking this is too little, too late. After a few years of playing around with VIA systems, I've moved to Intel boards where opensource drivers just work.
Nevertheless, it would be nice to see this work for real. Competition is always welcome. Frankly, it's weird that VIA hardware is geared towards embedded/mobile use, while providing drivers mainly for the most power-hungry OS on the planet.
Parent
I like VIA's strategy (Score:5, Informative)
VIA created the nanobook [via.com.tw] reference design for mini-laptops that use their low-cost, low-power chips. Already the CloudBook [wikipedia.org] has come out based on that design, and in other countries various similar laptops have been released from different distributors. Now they're stimulating essential linux development, which will continue to increase the value of their low-cost platform. This has "win" written all over it; we're all going to come out ahead thanks to their strategy.
Extend It To Crypto (Score:5, Informative)
Re: (Score:2)
Re:Extend It To Crypto (Score:5, Insightful)
Parent
Re:Extend It To Crypto - already done (Score:5, Interesting)
Via really stuffed up, however, when they made repeated half-baked attempts at 'semi-binary' drivers which worked only on ancient versions like Redhat 9 and wouldn't provide any support or information on the MPEG decoding chip (in the CLE266 and above) which was essential to getting working DVD and DVB playback on the low power boards like the Nemiah.
I won't hold my breath...
Parent
Re: (Score:2)
Via really stuffed up, however, when they made repeated half-baked attempts at 'semi-binary' drivers which worked only on ancient versions like Redhat 9 and wouldn't provide any support or information on the MPEG decoding chip (in the CLE266 and above) which was essential to getting working DVD and DVB playback on the low power boards like the Nemiah.
I agree with the general sentiments on the VIA-Linux relationship. However, my old system with a 1 GHz Nehemiah [sic] and CLE266 played DVDs just fine without MPEG2 acceleration, using MPlayer.
Cool (Score:4, Interesting)
VIA... open... 3d.... driver..... (Score:2)
VIA Envy24 Audio Chipsets (Score:3, Interesting)
will work with the community eh? (Score:3, Insightful)
Re: (Score:3, Insightful)
If they just publish the specs and stand back, I'm fine with that.
If they also release some of their current code under a free license it's a nice bonus.
ANOTHER bonus!
linux.via.com.tw -- no such host (Score:4, Insightful)
Their press announcement arrived before any content, sigh.
The PR says the website will start with: drivers, technical documentation, source code, and information regarding the VIA CN700, CX700/M, CN896 and the new VIA VX800 chipsets. It'd be good to see docs on their more widely used chipsets, like vt8235 and vt8237 ... detailed ones, including errata. I mean, currently they piss off almost everyone who uses their chipsets, so why would anyone want to buy NEW hardware with VIA chips if it's not even clear the current stuff can be made to work well?
It's a nice idea, years overdue. But even at that, pre-announced.
Re: (Score:3, Funny)