Any news on ATI fglrx 8.42?

[fatfatwolf]

I just can't wait to see this "legendary edition" with AIGLX support and 50% improvement on my X1300... I didn't even install 8.41 so that 8.42 would look like a huge brand new start for me.:D

Anyway, comparing to 8.41, it seems that 8.42 is really quiet. Anyone can give a confirmation that it actually supports AIGLX and R500s?

[Michael]

I am not allowed to talk about 8.42 until the day of the release or AMD permits it otherwise, but it has been stated before and in AMD's official press release that 8.42 should contain AIGLX support and is designed for pre-R600 graphics cards as well.

[fatfatwolf]

OMG, u just make me want to jump out of the window and rush for 3 hours so that I can exhaust myself and sleep all along until it gets released...

Waiting is painful...

[uhoh]

What's the deal with all the secret stuff? What do they stand to lose on early beta reports? Is it.... because they won't deliver again? :eek:

[Michael]

What's the deal with all the secret stuff? What do they stand to lose on early beta reports? Is it.... because they won't deliver again? :eek:

Relating to their Linux drivers I have a NonDisclosure Agreement that have been standing with them for over two years now, and unless that is nullified, I am not really able to talk about their future development plans.

[Svartalf]

What's the deal with all the secret stuff? What do they stand to lose on early beta reports? Is it.... because they won't deliver again? :eek:

Micahel's got an NDA with 'em. Means he CAN'T discuss anything about it. At the minimum, he'd get cut clean off from Beta Access. At the worst, they'd sue him for breach of that NDA.

Nobody here, Michael especially, wants either situation to come to pass.

I can wait. Lord only knows I've waited as long as I have for the silly thing- and it's still suboptimal in at least some ways. Perhaps 8.42 will do it. Heh...perhaps the Open Source team will beat 'em to a quality driver.

[mcgreg]

indeed there was a nice performance boost for x1300pro here :)

I just hope 8.42.x will have xorg 1.3 support and the 3D stuff fixed (the only reason why I dont use 8.41.7 is, I cant play diablo2 with it ;)

2.6.23-kernel support would be nice too, of course :)

[yoshi314]

I just can't wait to see this "legendary edition" with AIGLX support and 50% improvement on my X1300...your expectations are over-the-roof. i bet you'll be dissapointed.

i'm in no hurry for new fglrx. i'd rather wait a bit longer for radeonhd driver to mature.

[fatfatwolf]

your expectations are over-the-roof. i bet you'll be dissapointed.

i'm in no hurry for new fglrx. i'd rather wait a bit longer for radeonhd driver to mature.

Well, the ultimate goal is definitely radeonhd, but it'll just take too long. Meanwhile, I'll be really pleased if fglrx could get AIGLX running on my laptop.

[yoshi314]

Meanwhile, I'll be really pleased if fglrx could get AIGLX running on my laptop.make sure your eyes can take it. mine can't.

i guess i have a motion sickness of sorts, so i don't really care about compiz & stuff :(

[mits]

Any chance to get proper 2d in 8.42 (aiglx is just too much) if we sacrificed a dozen maidens? :D

[greg]

make sure your eyes can take it. mine can't.

i guess i have a motion sickness of sorts, so i don't really care about compiz & stuff :(
Just switch off wobbly windows, it doesn't have any use anyway. I think it's annoying too.

[enzobelmont]

any date hint???

[Michael]

any date hint???

Honestly, I don't even know right now... I would expect at the very end of the month.

[adam.tropics]

Honestly, I don't even know right now... I would expect at the very end of the month.

lol....playing it safe!

[yoshi314]

Just switch off wobbly windows, it doesn't have any use anyway. I think it's annoying too.even without wobbly and other wowie-zowie fx plugins my head and eyes hurt a lot when using it :/

i thought it was the refresh rate, but i don't use a crt monitor :/

[mile]

Honestly, I don't even know right now... I would expect at the very end of the month.

new bugs are poppin' out :)

[Kano]

Maybe on 31st they will deliver the "No AIGLX for ATI Edition" ;) NV started to support AIGLX only one year ago...

[pierluc]

Have-you test the new 8.42 driver?

[d2kx]

Have-you test the new 8.42 driver?

He could tell you, but he would have to kill you after that :D

Just kidding :p

[notz]

how about in xvideo support for RS690 chipsets?

it's a shame for amd/ati, that the the standard onboard chip can not play videos in linux since product launch (mor than 6 months ago)...

[Michael]

how about in xvideo support for RS690 chipsets?

The AMD Linux GPG group is and will be putting forward a lot of work with the RS690.

[notz]

The AMD Linux GPG group is and will be putting forward a lot of work with the RS690.

:-(, so no xivdeo support in the next release, too bad.

i hope the chipset will be supported by the radeonhd driver soon. but is there xvideo currently working with already supported cards?

[Michael]

:-(, so no xivdeo support in the next release, too bad.

i hope the chipset will be supported by the radeonhd driver soon. but is there xvideo currently working with already supported cards?

The RS690 should be supported already. AMD's video method of choice is TexturedVideo, which does use X-Video.

[bestis]

The RS690 should be supported already. AMD's video method of choice is TexturedVideo, which does use X-Video.

Should? Ok, tested today/yesterday with Centos5 and Fedora 7, with 8.41.7 and 8.40.4. Similar result always. Black screen and Xorg taking 100% of CPU.

I'm not very impressed in ATI drivers. This was my first go with them. Oh and luckily i haven't been able to get ANY graphical out of it. Not even with "vesa" driver.

Maybe tomorrow i will test radeonhd or avivo, if i want to play more, but i think i will reject the hardware and choose something that works out of the box.

[pierluc]

He could tell you, but he would have to kill you after that :D

Just kidding :p

If he give to me a copy of the driver, after he can kill me on a multiplayer video game.:D

[Michael]

Should? Ok, tested today/yesterday with Centos5 and Fedora 7, with 8.41.7 and 8.40.4. Similar result always. Black screen and Xorg taking 100% of CPU.

I'm not very impressed in ATI drivers. This was my first go with them. Oh and luckily i haven't been able to get ANY graphical out of it. Not even with "vesa" driver.

Maybe tomorrow i will test radeonhd or avivo, if i want to play more, but i think i will reject the hardware and choose something that works out of the box.

AMD has told me the RS690 is well supported and suggested I run some Linux tests. But I say "should" since I haven't gotten around to trying it out yet.

[notz]

The RS690 should be supported already. AMD's video method of choice is TexturedVideo, which does use X-Video.


that answer i got on the radeonhd mailing list:

RS690 is still very much on our todo list. So we can't support this yet.
RS690 hardware is available though, we just need to find the time.

[bestis]

AMD has told me the RS690 is well supported and suggested I run some Linux tests. But I say "should" since I haven't gotten around to trying it out yet.

Ok. Well it didn't seem like it and that RS690 name for it wasn't so obvious, lspci just speaks about X1200 and there wasn't X1200 listed in AMD pages. Readme told that it's supported but the download page for the driver doesn't even list it.

But my interests about testing/playing isn't high and i don't even know how long i will have the machine at my hands. I'm just doing compability tests to see which laptop we will choose.

Machine in question is Dell Latitude D631, it's intel counterpart (D630) worked out of the box with graphics:)

[notz]

The RS690 should be supported already. AMD's video method of choice is TexturedVideo, which does use X-Video.

did you mean fglrx or radeonhd?

on fglrx i get this (8.41.7):
(WW) fglrx(0): Textured Video is currently not supported on IGP hardware.

with latest radeonhd i get the driver loading but no xvideo extension is available.

[Michael]

fglrx. RadeonHD isn't that far along yet.

[notz]

fglrx. RadeonHD isn't that far along yet.

michael can verify the status of TexturedVideo with 8.42.7 or ati developers?

it would be a great thing, because i have bought this mobo for a htpc and can't really use it until now. and i didn't want to wait another 4 months to get my tv working...

[Extreme Coder]

I plan on buying an MSI MoBo, MSI K9AGM2-FIH, which has the ATI Radeon X1250 IGP, and I want to know, how well is it working with Linux? Is its (3D) performance similar to its Windows one? (under the latest driver) And would its 690G chipset from AMD cause any trouble?

Maybe you Michael know some more about this? ;)

[pierluc]

Maybe you Michael know some more about this?

But the driver are not on the web site of AMD and Michael can not say any thing before that.

[Kano]

The hope dies at last ;)

[pedepy]

im actually new to ati under linux and im a bit confused with all the different names / chips and all that .. i used to run intel gma 950 on an older laptop until something went wrong and i had to send the machine to warranty. Now then sent me this ati laptop, and im not gonna complain, considering all I got for virtually free.

i got an ati x1600 mobility (?) chip on my laptop and currently running xgl with the 8.40 driver; will my card be supported in 8.42 ? i have no tried out 8.41 as I understood it didnt support all types of card and 8.42 was supposed to offer broader support.

is that correct ?

[Michael]

im actually new to ati under linux and im a bit confused with all the different names / chips and all that .. i used to run intel gma 950 on an older laptop until something went wrong and i had to send the machine to warranty. Now then sent me this ati laptop, and im not gonna complain, considering all I got for virtually free.

i got an ati x1600 mobility (?) chip on my laptop and currently running xgl with the 8.40 driver; will my card be supported in 8.42 ? i have no tried out 8.41 as I understood it didnt support all types of card and 8.42 was supposed to offer broader support.

is that correct ?

Hi,

Welcome to the forums.

Yes, your card will be supported in 8.42.

[MagicTempest]

I would love to see the new driver. I hope it will support my x1950 AGP.

The 8.41 driver does give me a good performance boost with my old 9600Pro, but I would like to use my new x1950. At the moment that card is collecting dust in one of my drawers. :)

[Pexi]

Only one question, I have an ATI X1300 mobile, Will add ATI support for this chipset in the new driver release?

I hope so, because I'm impacient to try the new improvements and of course the combination of COMPIZ+AIGLX+other 3D software like google earth, enemy territory quake of wars,...

[Michael]

Only one question, I have an ATI X1300 mobile, Will add ATI support for this chipset in the new driver release?

I hope so, because I'm impacient to try the new improvements and of course the combination of COMPIZ+AIGLX+other 3D software like google earth, enemy territory quake of wars,...


Yes, that is an R500 chip.

[Torren]

I plan on buying an MSI MoBo, MSI K9AGM2-FIH, which has the ATI Radeon X1250 IGP, and I want to know, how well is it working with Linux?

I have it running here as a HTPC. 2D is running good, TV-OUT is fine. There is no video acceleration, so I use OpenGL instead which doesn't hurt picture quality.


Is its (3D) performance similar to its Windows one? (under the latest driver) And would its 690G chipset from AMD cause any trouble?


I don't know how well it performs under Windows, but 3D performance is ok. I can run World of Padman in 1024x768 with all settings set to above-medium (but not highest) and it runs with about 25 to 50 FPS. This is in a box with a Athlon64 X2 3800+ and 1GB of memory. Main problem is that the mem is only single channel at the moment, which really hurts IGP performance.

[Extreme Coder]

I have it running here as a HTPC. 2D is running good, TV-OUT is fine. There is no video acceleration, so I use OpenGL instead which doesn't hurt picture quality.



I don't know how well it performs under Windows, but 3D performance is ok. I can run World of Padman in 1024x768 with all settings set to above-medium (but not highest) and it runs with about 25 to 50 FPS. This is in a box with a Athlon64 X2 3800+ and 1GB of memory. Main problem is that the mem is only single channel at the moment, which really hurts IGP performance.
Thanks a lot for your help!

[pedepy]

This is neat: http://phoronix.com/forums/showthread.php?t=5582


you know, i wouldnt mind running alpha/beta quality software on my machine if it would help amd/ati iron out the kinks ..

They should realize that although we expect the same kind of quality everyone else does, most users of open source software dont usually mind testing out other people's work to help out. Now, with closed source drivers we wont be able to do anything about the actual code, but i dont mind handing out bug reports if it helps amd deliver a fine, performant aiglx enabled driver.

[d2kx]

you know, i wouldnt mind running alpha/beta quality software on my machine if it would help amd/ati iron out the kinks ..

They should realize that although we expect the same kind of quality everyone else does, most users of open source software dont usually mind testing out other people's work to help out. Now, with closed source drivers we wont be able to do anything about the actual code, but i dont mind handing out bug reports if it helps amd deliver a fine, performant aiglx enabled driver.

Every company, may it be nVidia, Codeweavers or AMD, is running beta programs. If you want to help improving the fglrx driver, I am sure you'll find a way.

[uhoh]

Micahel's got an NDA with 'em. Means he CAN'T discuss anything about it. At the minimum, he'd get cut clean off from Beta Access. At the worst, they'd sue him for breach of that NDA.

Nobody here, Michael especially, wants either situation to come to pass.


I know what an NDA is, the point of my question was "What do they have to gain by enforcing the NDAs so strictly". If Michael didn't censor the fglrx and opengl version numbers in his screen shots, would people buy less ATI cards?

Also, if they have beta code that works somewhat, IMHO they should post nightly beta releases (with whatever huge disclaimer necessary) so people could see for themselves that something is happening. Right now there's absolutely nothing that indicates ATI cards will ever get AIGLX support.

[Michael]

Also, if they have beta code that works somewhat, IMHO they should post nightly beta releases (with whatever huge disclaimer necessary) so people could see for themselves that something is happening. Right now there's absolutely nothing that indicates ATI cards will ever get AIGLX support.

While small, that all requires extra resources by AMD... They have a hard enough time right now just sanitizing specifications as they only have two people at times working on the specs, etc.

No matter how huge of a disclaimer they have, they no matter what have some level of responsibility and support by just publishing the driver in the public domain.

[HighHo]

I dont see any need for nightly builds of the proprietary driver. I do however think it would be good if there was an open beta phase, when possible. As an Example and without any knowledge of how there current betas work.
8.43 is released on 1/11/07.
8.44 (early) Closed Beta is released 10/11/07 with first set of bug reporting.
8.44 Open Beta is released 20/11/07 allows testing on a greater scale for any major issues that may occur on further testing.
8.44 Final released 1/12/07.

While the time frame for open beta to final is short, hopefully if any major issues that occur at the start of the beta for certain card users / distribution users can be resolved if time permits the changes needed.

Anyway just my thoughts of how ATI could change there beta program, which im sure they wont. :)

[d2kx]

Well, AMD releases so called hotfix Windows drivers, often after 2 weeks of a new WHQL release. These releases are somewhere "Open Betas" but without any help for AMD. They could do the same for Linux, but the Linux drivers have too less developers that it would be so important unless there's a new Kernel or Xorg (where a hotfix would be excellent).

[MagicTempest]

Don't forget the amount of work that gets in an open beta release. The people at AMD/ATI need a lot of time working through the various bug reports of people, like me, who don't know anything about driver development.
By issuing only a closed beta release they make sure that only people who know at least a considerable bit about what they're doing submit bug reports during beta. Therefore it's less time consuming to sort all the good bug reports from the bad and decide the course of action to resolve these bugs.

[Kano]

As usually all ATI drivers share the same problems + some newly introduced ones I would not think that they don't know the bugs, but don't how to fix em - like that Modeline bug which never was fixed but known for about a year. Also the usual statement to support new X/kernel versions only AFTER final releases of em is well known, that leads to a 1-2 month delay till you get the real fix - just by the way they handle this. NV needed some weeks for Xorg 7.3, but at the same time they added support for upcoming kernel 2.6.23. Do you think ATI does not know that this kernel will be out soon? Nobody else in the Linux field would expect an ABI change when there is RC9, but you can already "hear" in your virtual ear that this will be the "excuse" for not adding support for it. Sorry, but this is definitely a wrong way to develop drivers that should work on every current Linux system without forcing users to patch the drivers on their own - NV releases at least some OFFICIAL patches if needed to fix issues when they don't update the driver itself. ATI users don't have that possibility...

[metala]

your expectations are over-the-roof. i bet you'll be dissapointed.

i'm in no hurry for new fglrx. i'd rather wait a bit longer for radeonhd driver to mature.
You have to wait to the end of the year to have mature 2d support for radeonHD. Theya re currently working on 2d acceleration. And as it sounds 3d is more complex and needs more time.
What i have read is that radeonHD will not have all the things the proprietary driver will. AMD/ATI can not unveil all the information, cause there are agreements.

[Alistair]

You have to wait to the end of the year to have mature 2d support for radeonHD. Theya re currently working on 2d acceleration. And as it sounds 3d is more complex and needs more time.
What i have read is that radeonHD will not have all the things the proprietary driver will. AMD/ATI can not unveil all the information, cause there are agreements.

Its quite clear from my initial reading of the documentation release that there will be 'bits' that *are not* in the docs. These 'bits' will be related to third party Intellectual Property agreements that ATI/AMD is already tied to. What precisely these missing bits will affect in the way of 2d drivers and 3d drivers I'm not clear on -- one thing that appears clear to me is that certain hardware based locking schemes will not have their codebase revealed. So it is possible that in the future there will be video material available to the proprietary driver that ****might**** not be available to the opensource driver.
Considering the depth of skills that is already assigned to the radeonhd driver team, I have full expectation that the driver will mature -- perhaps not before june of next year, but eventually, to the state at which the vast majority of common users in linux will be more than satisfied with it.

I can clearly see also that the tvout capabilities may not be released, again due to 3rd party IP agreements. I happen to have an odd perspective on this as I'm working on a media project platform to deliver (cable) television signals to computers directly. Our hiccup is not the ability to put the medium on the (computer) screen, this is easily done, but rather to deliver it consistently to downstream television like devices. We're quite excited with the OpenSource Intel folks but we've hit our own snags on ATI --and equally- on Nvidia.
And in reality - the corporate attitude doesn't really include 'linux' as a priority, but the bulk of the team on the project are *definately* trying to keep linux in scope of 'just works'

[Ole-Martin Broz]

haha. . waiting is painfull.. what about me..

Im waiting for 2x gpu blocks for 2900 XT that ive ordered, waiting for the new quadcore from amd, waiting for some radiators, waiting for crysis, quakewars for linux, 8.42 driver/aiglx support, steam for linux, my G9 mice i ordered, waiting for my 4gb ocz pc4000 platinium ddr memory. yeah.. my list is enourmous.

well.. cant wait till 8.42, food the FEED up in FF, waiting for it to arrive, and im waiting for ubuntu gutsy aswell .!

[fatfatwolf]

And now it's Oct. 10th...

I am half dead because of the waiting...

[petercah]

Hi,

I have a thinkpad t60p with the mysterious v5250 (m56gl 71d4).. I am really struggling to find out for sure what exactly the architecture of this model is??

Is this card the same as the v5200/R520?? Will it be supported fully in the 8.42 release?? Could someone please clarify?? Would be much appreciated....

[Michael]

Hi,

I have a thinkpad t60p with the mysterious v5250 (m56gl 71d4).. I am really struggling to find out for sure what exactly the architecture of this model is??

Is this card the same as the v5200/R520?? Will it be supported fully in the 8.42 release?? Could someone please clarify?? Would be much appreciated....

Hi,

Yes, it's part of the R500 series and will be supported in 8.42.

[petercah]

Thanks.....that's music to my ears....so long XGL

[heruan]

Why AMD/ATI never announces release dates for its drivers?

[ivanovic]

Why AMD/ATI never announces release dates for its drivers?

Plain and simple said:

IIRWIIR (It Is Ready When It Is Ready)


Fixed release dates already quite some time before it can be out are something that is IMO even a bad thing. The problem is that you get a pressure of releasing that date even when there are grave bugs in the program. I prefer to wait some more days and have less bugs. That is one general thing why many open source programs do not have fixed release dates for *anything*, the really big ones just keep a general guideline that is nothing more than a target to meet but that will often not be met, what even is good because it means more time for new features and bugfixes.

Regarding the ATI drivers:
We do know that they want to release ~1 driver per month. My guess would be that the general aim is to release in the middle of the month to have some room for fixes if they are needed. If you think back to the last release you might remember that there were quite some bugs in it (hey, it was a complete rewrite of a subpart, that is normal) which have to be fixed now. Fixing bugs is not like "hey there is a bug, it will be dead in 5mins", fixing bugs often involves long time research for the root cause of the problem.
Since AMD/ATI wants to provide AIGLX with that new driver (at least regarding what Michael wrote in his reviews of 8.41) means that not only bugfixing has to happen but also implementing (and stabilizing) of it which takes quite some ressources.
I think a good working driver maybe around the 25th (very vague guess) is better than having a really buggy driver by the middle of the month simply because "a date was announced and we wanted to stick to it". Better give the devs the time needed to make a good product...

[Michael]

Since AMD/ATI wants to provide AIGLX with that new driver (at least regarding what Michael wrote in his reviews of 8.41) means that not only bugfixing has to happen but also implementing (and stabilizing) of it which takes quite some ressources.

Yeah, that remains. if you are prone to motion sickness be sure to have Compiz Fusion disabled this month!

[Namtabmai]

Hi, found this thread while looking for information on 8.42. Good to hear that the next driver will have AIGLX support!

I realise that due to NDAs that this question probably can't be answered, but...
Are ATI aware of the recent kernel change from SLAB->SLUB (I believe it has something to do with page allocation). While the current drivers work with this change for the most part, it does break suspend/hibernate with the drivers. There's a bug on Ubuntu LaunchPad (https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/121653) regarding this.

[d2kx]

Why AMD/ATI never announces release dates for its drivers?

Because AMD believes in the Duke Nukem Forever-release plan.

[Nicolas]

Hehehehe...

I'd rather get a late WORKING release than a timed worthless release.

Since I will be on vacation for three weeks (probably before 8.42's release), when I come back I hope I will have Ubuntu 7.10 + working FGLRX with AIGLX, so everything will be ok! =)

I haven't tried 8.41 because I have R300 hardware (Mobility X700). I chose to wait for 8.42's full support for my hardware and for AIGLX (if they can keep their promise).

Even though I'm still disappointed that they took so long to do something, I feel like they are at least going somewhere now. At least they are trying...

Good luck to all of us! :D

[Kano]

R300 is no problem for 8.41.7, just not all features are working and for wine you need an update. Major games should work with it.

[Nicolas]

Good to know... But I don't have much free time right now. When I get back from my vacation I'll update my Ubuntu to 7.10 (I'm still on 6.10) and get fglrx 8.42.

Should bring some nice improvements overall, and hopefully AIGLX!

[Raven3x7]

Well the windows driver just got released

[aromanos]

I haven't tried 8.41 because I have R300 hardware (Mobility X700).


Hey, the Mobility X700 is not a R300-class card, it's a proud R410 (a scaled down X800's R420) :D

Look at:
http://en.wikipedia.org/wiki/Radeon_R420
and
http://en.wikipedia.org/wiki/Comparison_of_ATI_Graphics_Processing_Units#Mobili ty_Radeon_Series

By the way, I also have a Mobility X700 and the 8.41.7 driver wasn't stable at all on my Ubuntu 7.04 + fglrx + XGL + Compiz.

Now I'm looking forward to Ubuntu 7.10 + fglrx + AIXGL + Compiz Fusion :)


-- Andrés

[pierluc]

Well the windows driver just got released (october 11)

Do you think that we will have the new driver Monday (october 14)? To drver of september had go out 2 days after the Windows driver.

[Kano]

The standard ati xserver seems to be more buggy with ubuntu than in debian for my X700 SE (which does not like a ati -> fglrx change without reboot or between some different fglrx drivers), compiz does not work with it - but compiz starts usally by default for this driver with ubuntu 7.10. The enabled features are however absolutely minimal in default settings so you would not even "see" that it runs - cube is disabled. For this card a new fglrx driver with AIGLX support might be a better choice. Let's see when it is done... I already gave Septor some patches for more generic X org detection to be used in U/D build scripts, lets see if they are included...

[HighHo]

New Windows drivers yesterday, im guessing that linux and windows drivers will no longer be released on the same date?

[Squizzle]

It would be great if this driver coincided with the final release of Gutsy Gibbon, I am using the current restricted driver in Fiesty (whichever driver version that is) on a laptop with the Xpress 1150 chipset... it's weak. desktop effects do not work without reverting to older versions of beryl and installing a diff X server etc. :( by the way... hi everyone :)

[Michael]

It would be great if this driver coincided with the final release of Gutsy Gibbon, I am using the current restricted driver in Fiesty (whichever driver version that is) on a laptop with the Xpress 1150 chipset... it's weak. desktop effects do not work without reverting to older versions of beryl and installing a diff X server etc. :( by the way... hi everyone :)

Hi,

Welcome to the forums.

It's an interesting thought having fglrx 8.42 coincide with the Gutsy Gibbon release. :cool:

[koolmanoncampus]

Why does it matter if its included in the ubuntu release? I'd rather have them wait for the driver to be good to release even if its oct. 31; then to rush delivery. Don't ubuntu package maintainers just put the update for the binary blob in the repository?
Besides can't u-users just install it with the download from ati?

Does anyone know what the state of fglrx, esp. in distros that push it like ubuntu, will be when RadeonHD is ready?

[Michael]

Why does it matter if its included in the ubuntu release? I'd rather have them wait for the driver to be good to release even if its oct. 31; then to rush delivery. Don't ubuntu package maintainers just put the update for the binary blob in the repository?
Besides can't u-users just install it with the download from ati?

Does anyone know what the state of fglrx, esp. in distros that push it like ubuntu, will be when RadeonHD is ready?

fglrx 8.41 or 8.42 is not being included by default in Gutsy.


RadeonHD will hit Hardy at the earliest.

[Nicolas]

Why does it matter if its included in the ubuntu release? I'd rather have them wait for the driver to be good to release even if its oct. 31; then to rush delivery. Don't ubuntu package maintainers just put the update for the binary blob in the repository?
Besides can't u-users just install it with the download from ati?

Does anyone know what the state of fglrx, esp. in distros that push it like ubuntu, will be when RadeonHD is ready?

It's not that it "matters". It just would be nice. I agree with you that waiting a few days for a good driver is the best option, but it would be good to have it released with Ubuntu 7.10.

[Nicolas]

fglrx 8.41 or 8.42 is not being included by default in Gutsy.


RadeonHD will hit Hardy at the earliest.

My guess would be that if 8.42 is stable and AIGLX capable, it might be available through repositories quickly?

As for RadeonHD, I don't have any idea. As an user, I would choose the best performing driver. When RadeonHD is as good as fglrx (functionality-wise) I might try it.

[Squizzle]

Why does it matter if its included in the ubuntu release? I'd rather have them wait for the driver to be good to release even if its oct. 31; then to rush delivery. Don't ubuntu package maintainers just put the update for the binary blob in the repository?
Besides can't u-users just install it with the download from ati?

Does anyone know what the state of fglrx, esp. in distros that push it like ubuntu, will be when RadeonHD is ready?

Ah, wasn't saying it mattered particularly, it's just that the new ubuntu enables compiz by default, I was merely making a passing thought that it would be nice if it all worked "out of the box" so to speak with my card when I upgrade on the 18th.

Couldn't agree more though, if it's the case, I'd much rather wait a few days for a decent driver than using a rushed one :)

Edit: Just read Michael's response to your comment and I must say I'm confused now.

I have used linux for some time, only most of it has been terminal stuff, administrating servers via PuTTY, having only switched at home a few days back, I'm new to all this video driver business... hence lurking here to try and learn a thing or two =)

[koolmanoncampus]

It's not that it "matters". It just would be nice. I agree with you that waiting a few days for a good driver is the best option, but it would be good to have it released with Ubuntu 7.10.


I guess I don't see the benefit. I wouldn't see why a linux distribution should focus on the release of proprietary software. If a user chooses to use them proprietary software I guess I don't see why they need their distribution to push it to them, rather than just get it themselves.

I am also surprised a linux distribution puts proprietary blobs in their repositories.

[pierluc]

It's more easy to install the driver fglrx of Ubuntu that to install the driver of the web site of ATI-AMD. But, the driver of Ubuntu is not the best and if ATI-AMD put the new driver on the web site before october 18 Ubuntu will be able to updates there driver to the 8.42.:)

[d2kx]

It's more easy to install the driver fglrx of Ubuntu that to install the driver of the web site of ATI-AMD. But, the driver of Ubuntu is not the best and if ATI-AMD put the new driver on the web site before october 18 Ubuntu will be able to updates there driver to the 8.42.:)

They won't do that.

[heruan]

They won't do that.

Did you mean "they" for "the Ubuntu team will not update the driver to 8.42" or "AMD will not release the driver before october 18"?

[Michael]

So guys, what else do you think is on the table for 8.42 aside from the pre-R600 fixes? :)

[d2kx]

Did you mean "they" for "the Ubuntu team will not update the driver to 8.42" or "AMD will not release the driver before october 18"?

I meant Ubuntu will not include them for Gutsy. There is a chance that the fglrx 8.42 gets backported for Gutsy but I would not trust in this. However, you can, of course, install the fglrx by yourself.

So guys, what else do you think is on the table for 8.42 aside from the pre-R600 fixes?

You've spoken about AIGLX support in recent articles, so I guess it should be included. You've mentioned that there is no Kernel 2.6.23 support, which is a bit disappointing but maybe we can have Xorg 7.3. Oh, and the fglrx 8.41 did not have 3D options in the AMDCCCLE. I pray that these will come back.

[yoshi314]

So guys, what else do you think is on the table for 8.42 aside from the pre-R600 fixes?not much ? :D

oh, maybe new x.org support will arrive.

[heruan]

I meant Ubuntu will not include them for Gutsy. There is a chance that the fglrx 8.42 gets backported for Gutsy but I would not trust in this. However, you can, of course, install the fglrx by yourself.

In feisty (released on April, 19) there's fglrx 8.37.6 (released on May, 31). So I guess Ubuntu will probably include 8.42 in backports asap, especially if AIGLX support will be supported...

[Malikith]

So guys, what else do you think is on the table for 8.42 aside from the pre-R600 fixes? :)

Probably anti aliasing support for HD 2xxx series cards. Some bug fixes (if we're lucky), compiz, the new xorg support. And if we're luckier, maybe some fixes for xv video playback. Maybe enhance some of the 2d performance a bit which would help the xv video playback.

Maybe umm... Ohh wait, no they won't do that (fix agp bus issue on x1k cards). If they do, i'm sold, that shouldn't be a hard bug to fix anyway, its a detection problem.

[kkloster]

I'm new here so sorry if this is a silly question: will the 8.42 fglrx driver bring support for hardware acceleration of decoding h.264 video or VC-1 (taking advantage of the UVD)?

that isn't already supported, is it?

[yoshi314]

I'm new here so sorry if this is a silly question: will the 8.42 fglrx driver bring support for hardware acceleration of decoding h.264 video or VC-1 (taking advantage of the UVD)?i guess not. maybe it's implemented in textured video feature of fglrx, maybe not.

[Fenix-TX]

When will fglrx driver fixes bug with a capture tv? I've always have problem with fglrx and capture tv. If o open any tv program, like kdetv, xawtv, ... the screen switch to black (in the best case) or crashes x server (the worse case).

[ivanovic]

When will fglrx driver fixes bug with a capture tv? I've always have problem with fglrx and capture tv. If o open any tv program, like kdetv, xawtv, ... the screen switch to black (in the best case) or crashes x server (the worse case).

I don't really understand what you want to do exactly. I am able to watch TV using tvtime (analog TV via cable) and the fglrx driver. Those are the relevant parts of my config for this stuff:

Section "Extensions"
Option "Composite" "disable"
# fglrx-only für xvideo (tvtime)
Option "XVideo" "Enable"
EndSection

Section "Device"
Identifier "ATI Graphics Adapter"
Driver "fglrx"
BusID "PCI:1:0:0"
Screen 0
Option "MonitorLayout" "TMDS, NONE"
Option "AGPMode" "8"
Option "RenderAccel" "on"
Option "AccelMethod" "xaa"
Option "GARTSize" "64"
# noetig für flüssiges Bild unter tvtime
Option "XAANoOffscreenPixmaps" "true"
# r300 driver options
# Option "EnablePageFlip" "true"
# Option "ColorTiling" "true"
# fglrx options
Option "VideoOverlay" "on"
Option "OpenGLOverlay" "off"
Option "UseInternalAGPGART" "no"
EndSection


As you see I do have VideoOverlay set to "on" and "OpenGLOverlay" to off. I do need VideoOverlay active to have tvtime work, but that is all (beside the xvideo extension which is not needed when using the open source radeon driver).

In short: watching TV is possible over here using the fglrx driver though the output does not look as crisp as with the open radeon driver. At least when not running at full screen it seems to be rather "blocky" especially with text. When capturing a screenshot from inside tvtime that is not oberservable (as well it is working nicely with the open driver) so it is definatly an issue with the xvideo implementation in the fglrx driver.

PS: I was never able to get things working using kdetv but I do at least get an image using xawtv. My graphics card is a ATI Radeon 9800Pro and my TV-Card is a Cinergy 400 TV (using the "Philips Semiconductors SAA7134/SAA7135HL Video Broadcast Decoder (rev 01)
" (regadring to lspci)). I had initially some problems to get the tuner working since autodetection does not work nicely...

[Fenix-TX]

My problem that i've always had with fglrx and not with radeon driver, is when i open a program that uses bttv (tv capture), sometimes x server crashes (very frequent), and other times it turns black the screen, so i have to switch to a virtual terminal (ctrl+alt+(F1-F6)) and back again to desktop (ctrl+alt+f7).

I've seen this problem searching in google with a lot of people with fgrlx and programs that uses tv capture. All the answers says it's a fglrx problem. With radeon driver, i've never had problems with my tv capture and using tv programs.

EDITED: This is my configuration:
Section "Device"
Identifier "ATI Technologies Inc RV350 AQ [Radeon 9600]"
Driver "fglrx"
BusID "PCI:1:0:0"
Option "VideoOverlay" "on"
Option "OpenGLOverlay" "off"
Option "XaaNoOffscreenPixmaps"
Option "UseFastTLS" "2"
Option "BusType" "AGP"
Option "PseudoColorVisuals" "off"
Option "UseInternalAGPGART" "no"
Option "mtrr" "off"
Option "no_accel" "no"
Option "no_dri" "no"
Option "EnablePrivateBackZ" "no"
EndSection

[Nicolas]

So guys, what else do you think is on the table for 8.42 aside from the pre-R600 fixes? :)

Besides pre-R600 support and AIGLX? I wouldn't ask much... If they deliver that I'm happy already.

Hopefully we're near the release date now, so we shouldn't have to wait long to use it for real :D.

[kkloster]

sorry to repeat this but I realized that yoshi314 and rthorntn actually discussed it in a different thread and there didn't seem to be a solid answer so maybe Michael or someone can clear this up:

is UVD currently supported in fglrx 8.41 ?

if not, is it a possibility that 8.42 will add support for hardware decoding h.264, VC-1 (i.e, UVD support)?

[Ole-Martin Broz]

AIGLX should be supported from r400 cards aswell, from there on :p
cause there is not many people who using pre r400 for games aswell.

[Malikith]

I hope today is the day the 8.42 driver comes out. I'm pretty anxious to see what it brings to the table.

[Phlogiston]

I hope today is the day the 8.42 driver comes out. I'm pretty anxious to see what it brings to the table.

I hope so too, it is not out yet.

[Michael]

I hope so too, it is not out yet.

Don't expect it today.

[pierluc]

Micheal, do you know the date when the driver will go out? If you know that it's not today I think it's because you have the real date.

[Michael]

Micheal, do you know the date when the driver will go out? If you know that it's not today I think it's because you have the real date.

I am afraid I can't comment on specific dates due to NDA.

[pierluc]

So you know the date because if you don't know the date you can said "no".

Ok, I will Wait.

Will you give a test of Catalyst 8.42 in the same time that ATI put the new driver on the web?