RadeonHD Driver Takes A Blow In Novell Layoffs [Archive]

duby229

03-02-2009, 09:54 AM

Well, it's sad to see such a prominent developer leave the community. I would hope that he would choose to continue his work, at least on the development side.

Otherwise I think this is just one more example of where RadeonHD went wrong. To rely on a single corporation is just suicidal. Especially when that single corporation is out to get you. I knew from the beginning RadeonHD was doomed. It relied exclusively on Novell. And it is a fact that Novell is out to screw the open source community.

Hopefully we can all use this as an excuse to --FINALLY-- consolidate the development of an ATi X driver onto one project. Instead of having one driver that's good with this thing, and another that is good with that thing. Lets get it all consolidated into one project that is good with all things....

Michael

03-02-2009, 09:57 AM

Well, it's sad to see such a prominent developer leave the community. I would hope that he would choose to continue his work, at least on the development side.

From what he has said to me, don't count on it. Unless he ends up working for another major corporation as an X developer.

tuxdriver

03-02-2009, 10:04 AM

I hope he continues his work on the proper Radeon driver.

Due to limited access now to ATI hardware, it's unlikely that we will see any future work from him in the RadeonHD driver.

We can always donate ATI hardware and/or cash if he needs it :)

yoshi314

03-02-2009, 10:04 AM

From what he has said to me, don't count on it. Unless he ends up working for another major corporation as an X developer. even nowadays, such an experienced developer should not have big trouble finding a new job.

some company is bound to hire him sooner or later.

aaaantoine

03-02-2009, 10:10 AM

even nowadays, such an experienced developer should not have big trouble finding a new job.

some company is bound to hire him sooner or later.

Yes, especially given his relative fame.

Vadi

03-02-2009, 10:15 AM

Didn't see X.Org development mentioned explicitly, but since Canonical has the Bryce guy who's involved, maybe they'd be interested: http://webapps.ubuntu.com/employment/

rbmorse

03-02-2009, 10:57 AM

even nowadays, such an experienced developer should not have big trouble finding a new job.

some company is bound to hire him sooner or later.

He won't have much trouble finding a new job, but it's highly unlikely it will be one where he's paid to develop an open-source Linux driver for ATI chipsets

deanjo

03-02-2009, 11:02 AM

Wouldn't surprise me if Apple or nv snaps him up. Of course ATI/AMD should be the first to grab him.

Louise

03-02-2009, 11:24 AM

Wouldn't surprise me if Apple or nv snaps him up. Of course ATI/AMD should be the first to grab him.

Or Red Hat :)

Redeeman

03-02-2009, 11:49 AM

I would hope that hardware isnt a factor in him discontinuing development - as have already been mentioned, i doubt it would be a problem to get hardware donations - i would personally gladly contribute some hardware or money to buy the hardware, should he decide to keep developing.

monraaf

03-02-2009, 01:20 PM

Sad, but I'm sure he'll find another job.

What I don't get is what the Novell developers are actually contributing to RadeonHD. I've been tracking the progress of RadeonHD for a couple of months now, and to me it looks like 99% of the code has been written by Alex Deucher.

Chewi

03-02-2009, 02:20 PM

Ouch. I had a feeling something like this might happen. Thanks for your hard work, Luc.

bugmenot

03-02-2009, 02:47 PM

Oh my god. Now it's getting even harder for radeonhd to exist parallel to radeon. Thanks for your great work in so many projects! Hope you find a new job soon.

Arch64

03-02-2009, 02:49 PM

I hope he goes to help xf86-video-ati in their r600-r700 code.

grantek

03-02-2009, 05:52 PM

Yeah, hopefully one of the big OSS companies will pick him up, or if he wants a direction change in his career, hopefully he'll still be around to advise/mentor new devs :)

cruiseoveride

03-02-2009, 09:19 PM

well it was fun to watch the radeonhd driver get this far

Saist

03-02-2009, 09:44 PM

I'm not sure the story is "over" for the RadeonHD driver. The information used by the developers is still available, and unless things have changed where AMD's not releasing any more information (a job headed internally by Mr. Bridgman I believe), the code documentation to build RadeonHD is still around.

***

I did however have an off-shoot idea. Via's been trying, albeit poorly, to establish a good direction in Open-Source development. I wonder if Via's thinking about hiring Luc on an official basis...

sreyan

03-02-2009, 10:28 PM

It's really a shame since radeonHD was so good and stable at what it was able to achieve. Best of luck Luc Verhaegen. As others have stated cash/and or hardware if you want it.

deanjo

03-02-2009, 10:47 PM

It's really a shame since radeonHD was so good and stable at what it was able to achieve. Best of luck Luc Verhaegen. As others have stated cash/and or hardware if you want it.

Question being why doesn't ATI just send out the hardware under NDA to such developers with a proven track record? They send tonnes of pre-release hardware to review sites why can't the same be done for the devels?

Saist

03-02-2009, 11:53 PM

Question being why doesn't ATI just send out the hardware under NDA to such developers with a proven track record? They send tonnes of pre-release hardware to review sites why can't the same be done for the devels?

also with Mr. Bridgman and Alex on staff, it's not like AMD doesn't have people who can TELL them who the good driver developers are, and who should be shipped some hardware.

***

I will point out though that most of the pre-release hardware sent to review sights actually comes from 3rd parties, e.g. Sapphire.

rbmorse

03-03-2009, 12:21 AM

If only they had a budget for that kind of thing....

Beancounter: "Let me get this straight...you want money to send some stuff to a guy who used to work for somebody else but got laid off, has no contract with us to do anything, in hope he might work for no salary on a low-visibility project that is costing us a great deal and has absolutely no chance of bringing in revenue in the future?"

Bridgeman: "Yes, that's pretty much it."

Beancounter: "I just came up with a brillinat idea of how we can save even more money. Clean out your desk."

Kjella

03-03-2009, 03:11 AM

Beancounter: "Let me get this straight...you want money to [spend money] on a low-visibility project that is costing us a great deal and has absolutely no chance of bringing in revenue in the future?"

The way you put it, there wouldn't be any reason to spend any money whatsoever. Obviously AMD is spending a lot of Bridgman and his team's time, lawyer time, executive time running things up and down the flag pole for release and so on. Compare that to say 10x $400+$100 for shipping and handling to the top 10 developers and it's a bargain. There might be a lot of other practical matters that means it's not possible but pure cost couldn't possibly be one of them.

Saist

03-03-2009, 04:04 AM

has absolutely no chance of bringing in revenue in the future?"

Actually, this isn't correct. The potential revenue is actually huge. AMD / ATi makes money off of chips that are sold, not off of selling software. Their goal, as a company, is to get the most chips into the market that they can.

An Open-Sourced driver means there are more platforms the chip could be supported.

Take the R200, or Radeon 8500 series for example. Imagine for a second if somebody ordered a bunch of R200 chips from AMD for a series of netbooks, low-end laptops, or so on. On today's manufacturing process's, the chip could be a lot smaller, and cheaper. TechArp notes that the Radeon 8500 was built on a .15 micron process with around 60million transistors. Today's lowest end RadeonHD, the 4350, has around 242 million transistors on a 55nm process, so it's chip made almost 3 times smaller than the 8500, while being several times more complex.

While AMD no longer supports this chipset internally with Fgrlx, that chip IS well supported under the X.org ATi, and is quite capable of driving tear-free video acceleration, 3D effects, and so on. An R200 built on a 55nm process would be astoundingly small, and could quite possibly run without any direct cooling.

Okay, I'll admit this is an extreme case, but it's just one way AMD could use their existing product portfolio to turn a profit with an Open-Source development process on the driver side.

Having an Open-Licensed driver also makes it easier to sell to OEM's, such as Dell. Remember the Dell Ubuntu systems launched with Nvidia due to Nvidia's reputation. However, after improvements to the ATi driver options, Dell started offering systems with ATi chips : http://www.phoronix.com/scan.php?page=news_item&px=NjYwMQ

which is AMD's goal... to sell more chips.

In my view, any accountant who can't wrap their heads around the business model AMD works on, and can't see how open-source can benefit that model, doesn't need to be working for AMD.

mirza

03-03-2009, 07:12 AM

jonnycat26

03-03-2009, 07:35 AM

And it is a fact that Novell is out to screw the open source community.

Oh yes, Novell certainly are a crafty bunch of miscreants. They contribute to the kernel, actually, they're they second biggest contributor behind RedHat and they're ahead of IBM.

What dastardly deeds could they be planning?

They contribute to KDE and still have 2 developers working on RadeonHD.

Pricks!

They're sure out to screw the open source community!

Svartalf

03-03-2009, 08:39 AM

Question being why doesn't ATI just send out the hardware under NDA to such developers with a proven track record? They send tonnes of pre-release hardware to review sites why can't the same be done for the devels?

It's believable. And it's not unheard of to get such hardware out of them (I've done it before in the past trying to verify LGP stuff- it's still sitting in my antiquated G4 PowerMac.)

Svartalf

03-03-2009, 08:45 AM

They're sure out to screw the open source community!

Actually, I don't think they're out to screw the community, per se. But with the recent remarks they've made about Linux, the pact they signed with MS, and the risks they take doing things for "interoperability" with Windows stuff (i.e. Mono...) by doing stuff that is very likely to be patented (and while Bilski will eventually clean out most of the risks in question- the bulk of their portfolio is on protocols, and other similar software "patents" which CAN'T be patented per Bilski...) which means they're opening up anyone using it to the risk of something like TomTom is now facing...they're not quite sane either.

They're not our friend like you make them out to be.

jonnycat26

03-03-2009, 08:57 AM

Actually, I don't think they're out to screw the community, per se. But with the recent remarks they've made about Linux, the pact they signed with MS, and the risks they take doing things for "interoperability" with Windows stuff (i.e. Mono...) by doing stuff that is very likely to be patented (and while Bilski will eventually clean out most of the risks in question- the bulk of their portfolio is on protocols, and other similar software "patents" which CAN'T be patented per Bilski...) which means they're opening up anyone using it to the risk of something like TomTom is now facing...they're not quite sane either.

They're not our friend like you make them out to be.

They're not our enemy like you make them out to be.

Mono is the pet project of Miguel, and I really don't get his reasoning behind it. Then again, I don't get gnome either. :)

Didn't redhat sign a similar deal with Microsoft? Are they the enemy too?

Fact is, Novell has been very good to Linux. I know, for a fact, that they've pushed Linux into places/companies where RedHat has failed.

Edit: BTW -- If you don't like Novell, you could always, ya know, just remove all of the Novell contributed code from the Kernel/QT/KDE. :)

NeoBrain

03-03-2009, 09:00 AM

BTW: Anyone else surprised by Nouveau progressing at same pace as AMD open source drivers ? At least as far as I can tell.

wtf? They are nowhere near to any usuable 3d acceleration and you call their progress comparable to that of radeon/hd?
They don't have reliable VT switching either... they're basically same as nv but with 2D acceleration. Apart from the general fact that the driver code for radeon/hd is probably /much/ cleaner.

Svartalf

03-03-2009, 06:16 PM

They're not our enemy like you make them out to be.

Oh, I didn't make them out to be an enemy. They're very seriously confused and aren't helping things any.

Mono is the pet project of Miguel, and I really don't get his reasoning behind it. Then again, I don't get gnome either. :)

If you don't get it, why do you diminish the whole thing? It's a problem, just like the deal was. If there's patent overlap, it's certain within the .Net stuff. Why bring that upon yourself?

Didn't redhat sign a similar deal with Microsoft? Are they the enemy too?

No. Red Hat signed something that talked to interoperability, but did NOT discuss patents or licensing thereof. It was a, "We're going to ensure that our Virtualization solutions are able to run each other's OS right" deal. Novell signed a deal where they indicated that there was valid patents within Linux that needed licensing from MS- which isn't a foregone conclusion, nor is it helpful when MS will not explicitly indicate what 200+ patents are infringed upon so that a discussion can ensue on the subject.

Fact is, Novell has been very good to Linux. I know, for a fact, that they've pushed Linux into places/companies where RedHat has failed.

They've also been very harmful to it with recent actions. I know that is also a fact.

Please, quit making them out to be saints (which is what you're actually doing here...)- I'm not vilifying them, mind, but they've done a lot of VERY problematic things over the years. Good deeds are important, but they do NOT remove bad deeds done in recent times.

deanjo

03-03-2009, 06:47 PM

Novell signed a deal where they indicated that there was valid patents within Linux that needed licensing from MS- which isn't a foregone conclusion, nor is it helpful when MS will not explicitly indicate what 200+ patents are infringed upon so that a discussion can ensue on the subject.

The agreement has no such thing. The agreement protects any future action that maybe brought by MS but does not admit to infringing any patents at all. You wouldn't say a fire detector is a sure sign of a impending fire now would you?

jonnycat26

03-03-2009, 07:10 PM

Please, quit making them out to be saints (which is what you're actually doing here...)- I'm not vilifying them, mind, but they've done a lot of VERY problematic things over the years. Good deeds are important, but they do NOT remove bad deeds done in recent times.

I'm not making them out to be saints, but they're certainly not out to destroy open source. And they actually contribute more than some very popular distributions U might have heard of.

Novell, tbh, has done far more good than harm. To say anything else is seriously short-sighted.

Anato

03-03-2009, 07:13 PM

The agreement has no such thing. The agreement protects any future action that maybe brought by MS but does not admit to infringing any patents at all. You wouldn't say a fire detector is a sure sign of a impending fire now would you?

Yeah, but it also limits Novell's ability to protect the community. And by singning it implies that there are some valid (as from Novell's point of view) concerns that MS might be hold such "patents".

Svartalf

03-04-2009, 12:06 AM

The agreement has no such thing. The agreement protects any future action that maybe brought by MS but does not admit to infringing any patents at all. You wouldn't say a fire detector is a sure sign of a impending fire now would you?

Only with their own customers and not the community as a whole...

You or I are NOT protected in that little deal. It's not a fire detector or a firehose. Bad analogy deanjo.

deanjo

03-04-2009, 07:29 AM

Only with their own customers and not the community as a whole...

You or I are NOT protected in that little deal. It's not a fire detector or a firehose. Bad analogy deanjo.

A fire detector only protects the area of the coverage. Not the entire city. Nor do I buy liability insurance with the intention of screwing up someone else's wellfare.

makomk

03-04-2009, 07:57 AM

wtf? They are nowhere near to any usuable 3d acceleration and you call their progress comparable to that of radeon/hd?
They don't have reliable VT switching either... they're basically same as nv but with 2D acceleration. Apart from the general fact that the driver code for radeon/hd is probably /much/ cleaner.

Well, the same as nv but with very good 2D acceleration on everything earlier than the 8000 series. The current-gen hardware is harder, since they can't use the 3D engine to draw to linear surfaces and haven't figured out how to make the framebuffer tiled yet. Oh, and I think they've had 3D mostly working on NV40-class hardware (equivalent of r500) for quite a while, and are now looking at doing stuff like LLVM on it.

(At least some of the issues with 3D seemed to be Gallium-related. Until very recently, nouveau was the only open-source driver for real 3D hardware that was using Gallium. The other one was the Intel driver, but I'm not counting that for obvious reasons. I don't think the Intel hardware that's targeting even has vertex shaders.)

Radeon/RadeonHD have, thankfully, caught up in most areas now. For quite a while after RadeonHD started, though, Nouveau actually seemed to be under heavier development than RadeonHD. It also had a head start in 2D acceleration, with fast 2D and Xv support for the nv40 cards back when RadeonHD didn't even exist and documentation from ATI was a pipe dream.

sreyan

03-04-2009, 09:39 PM

Novell has enough money for many developers of duplicated non-functional crap called "Mono" but not for basic HW drivers?

BTW: Anyone else surprised by Nouveau progressing at same pace as AMD open source drivers ? At least as far as I can tell.

Mono is the pet project of Miguel, and I really don't get his reasoning behind it.

Having been employed largely in the windows world it pains me to admit that .NET is actually really nice and slick; I enjoy programming in C# much more than I enjoy Java or C++.

I have actually used mono, and a client was impressed that we could run some server side software on linux instead of windows machines.

Mono certainly isn't perfect; I've personally found embarrassing bugs (see http://code.google.com/p/flowlib/issues/detail?id=10&can=1), but it does make cross platform development cheaper for small firms.

I also appreciate Novell's work supporting KDE, evolution, etc -- even if I don't use them

Compatibility is good. Competition is good.

duby229

03-04-2009, 11:16 PM

Having been employed largely in the windows world it pains me to admit that .NET is actually really nice and slick; I enjoy programming in C# much more than I enjoy Java or C++.

I have actually used mono, and a client was impressed that we could run some server side software on linux instead of windows machines.

Mono certainly isn't perfect; I've personally found embarrassing bugs (see http://code.google.com/p/flowlib/issues/detail?id=10&can=1), but it does make cross platform development cheaper for small firms.

I also appreciate Novell's work supporting KDE, evolution, etc -- even if I don't use them

Compatibility is good. Competition is good.

The problem is that Novell sold out the open source world to get it. I'd like to see what these 200 patents are.... How many of those patents are broken by Novell in projects like Mono, which they conveniently pass the blame to the open source world as a whole.

They infringe on other companies patents, then they dont tell us what those patents are, and then actually have the gall to blame the rest of us for it.... I mean jeez...

If you dont think that is a coordinated plan to ruin open source then your blind.

sreyan

03-05-2009, 09:42 AM

The problem is that Novell sold out the open source world to get it. I'd like to see what these 200 patents are.... How many of those patents are broken by Novell in projects like Mono, which they conveniently pass the blame to the open source world as a whole.

They infringe on other companies patents, then they dont tell us what those patents are, and then actually have the gall to blame the rest of us for it.... I mean jeez...

If you dont think that is a coordinated plan to ruin open source then your blind.

Of course we'd all like to know the patents in question, since In re Bilski probably makes many of them worthless. Given that, the actual numberer of real patents that free software users can be threatened with is probably an order of magnitude smaller.

I recall at one point a bunch of big corporations agreed to use their patent portfolio as a deterrent. Other than MS suing some gps manufacturer, I don't think you have anything to worry about.

duby229

03-05-2009, 05:23 PM

Of course we'd all like to know the patents in question, since In re Bilski probably makes many of them worthless. Given that, the actual numberer of real patents that free software users can be threatened with is probably an order of magnitude smaller.

I recall at one point a bunch of big corporations agreed to use their patent portfolio as a deterrent. Other than MS suing some gps manufacturer, I don't think you have anything to worry about.

I've done some reading on Bilski, but I cant quite grasp what the implications are. I'm not a patent lawyer and unfortunately I dont understand much of the terminology being thrown around in the blogs. Is there an english interpretation in laymans terms that you can point me to?

sreyan

03-05-2009, 05:43 PM

I've done some reading on Bilski, but I cant quite grasp what the implications are. I'm not a patent lawyer and unfortunately I dont understand much of the terminology being thrown around in the blogs. Is there an english interpretation in laymans terms that you can point me to?

Groklaw has tons on bilski, but the article that summarizes what it means in terms of MS's patent portfolio is here:
http://www.groklaw.net/article.php?story=20081102011538422