Archiv verlassen und diese Seite im Standarddesign anzeigen : Ati Linux Chatlog

2004-11-12, 00:16:21
Für die, die es verpasst haben. Hier ist das Chatlog vom Rage3D chat mit Ati.
ATI_Linux_Guys und mtippett sind von Ati.

[21:58:22] •››››››››››››››››››››››››››››››››››››››››››› ››››› ››› ›› ›› ›› › ›
[21:58:23] [@[R3D]ben6]: hi terry and matthew and Jon! :)
[21:58:32] [@[R3D]Ratchet]: ATI_Linux_Guys is Jon and Terry
[21:58:48] [@[R3D]dementor]: Don't forget to ask questions in #ATILinuxQ so that ATi may have the oppurtunity to answer it. We'll be posting questions based on their importance to the topic at hand (LINUX).
[21:58:49] [@[R3D]Ratchet]: mtippett is Matt, of course :)
[21:59:28] ::: Mode: ([R3D]dementor) sets (+D)
[22:01:24] [+ATI_Linux_Guys]: Hi everyone
[22:01:28] [+ATI_Linux_Guys]: ATI is here
[22:01:31] : +mtippett waves to the gallery
[22:01:37] [+ATI_Linux_Guys]: welcome to the 4th Rage3D chat
[22:01:45] [+ATI_Linux_Guys]: topic is Linux
[22:01:46] [@[R3D]ben6]: 5th terry
[22:01:49] [@[R3D]MrB]: 4:02PM EST....we'll start the questions now....we have a bunch of questions already logged by Ratchet that he deemed relevant.
[22:01:55] [+ATI_Linux_Guys]: ok Ben I stand corrected
[22:02:02] [+ATI_Linux_Guys]: 5th Rage3D exlusive chat
[22:02:11] [@[R3D]Ratchet]: ok, 1st question coming up
[22:02:19] [@[R3D]Ratchet]: <HWolf> Question: Anandtech recently did a comparison of linux performance between ATI and Nvidia cards. They reported they needed several kernel-recompiles, a lot of tweaking, and general problems. They ended up with inferior performance for ATI products at all fronts, aswell as being unable to run many of the benchmarks, what is ATI's responce to that?
[22:02:23] [+ATI_Linux_Guys]: we will type stop when the answer is done
[22:02:29] [@[R3D]Ratchet]: rgr
[22:03:44] [+ATI_Linux_Guys]: Matthew is typing a response
[22:03:56] [@[R3D]MrB]: jeopardy music in the backgroun.....
[22:04:02] [+ATI_Linux_Guys]: by the way Matthew is ATI's Project Team Lead for Linux
[22:04:22] [+ATI_Linux_Guys]: we also have Terry and Jon (CATALYST marketing and PR respectively)
[22:05:07] [@[R3D]MrB]: So Joh and Terry what do you guys think of SLI
[22:05:21] [@[R3D]MrB]: Jon
[22:05:41] [+ATI_Linux_Guys]: The Linux version of it you mean? ;-)
[22:06:12] [@[R3D]MrB]: we'll I think we need filler well Matt defends his linux drivers .... SLI is a hot topic
[22:06:12] [+mtippett]: Our current focus is on stability and features. We are getting incremental performance increases as they become available. So it was no surprise about the performance in the review. Our drivers are primarily targetted at Red Hat, and we work with SuSE to repackage our drivers, and so AnandTech got off to the wrong start by not using those. Once the SuSE packages where used, the issues about tweaking and general problems were not present.
[22:06:51] [+mtippett]: stop
[22:07:01] [@[R3D]Ratchet]: <skirst> QUESTION: GLSL has been in the Windows CATALYST OpenGL for about a year now, while it is currently abscent in the Linux OpenGL implementation. This suggests to me that the two do not share the same codebase. Are there plans to take a more unified approach so that developers can gain access to new OpenGL improvements and extensions at roughly the same time for either platform?
[22:08:02] ::: Mode: (ChanServ) sets (+o ben6)
[22:08:32] [+mtippett]: Since 3.9, the code has been tracking the windows codebase for each release. In upcoming releases we will match the Windows version number (December will be a 8.08 release).
[22:09:56] [+mtippett]: A feature disparity may in some cases (GLSL for example) may be absent since the code is not stabilised or there are technology restrictions for Linux. As we have opportunity we will bring features on par with windows.
[22:10:01] [+mtippett]: stop
[22:10:13] [@[R3D]Ratchet]: <oliver> QUESTION: Some voulenteers are starting to work on a nice GTK based control panel. Is there any way they(we) could get some detailed information or some api (the docs under NDA if wanted) to further work on that panel? Perferably via a direct channel and not the developer program on the ATi site that gets you no responce?
[22:12:12] ::: Nick: (ben6) is now known as ([r3d]ben6)
[22:12:31] [+mtippett]: Since our control panel is currently QT, we have sources shipped. We are working through some licensing issues in accepting patches. We have accepted patches from various people since I made the comment 'that we hadn't received anything for the control panel'.
[22:13:05] [@[r3d]ben6]: QT?
[22:14:06] [+ATI_Linux_Guys]: QT = underlying windowing library that KDE uses
[22:14:08] [+mtippett]: An idea that I am playing with (there are a few technical prerequisites before we can do this). Is that the control panel is developed in part by ATI and in part by the community - Rage3DTweaker is an example of some of the things that may be exposed.
[22:14:08] [+ATI_Linux_Guys]: sorry
[22:14:14] [@[r3d]ben6]: thanks terry :)
[22:15:18] [+mtippett]: There may be some ATI only extensions that will be included in the released ones in binary forms, but I have no fundamental problems with working with the user community to build up channel.
[22:15:51] [+mtippett]: On the QT vs GTK, currently we have some code in QT, and so GTK would be a reset in the control panel.
[22:15:56] [+mtippett]: stop
[22:16:06] [@[R3D]Ratchet]: <HWolf> Question: From your answers it looks to be a priority for ATI to get support done right on the 'business' linux platforms, especially since you at the moment only seem to focus on the RedHat/SuSe enviornments. What can the average enthouiast expect from ATI in the near future?
[22:17:34] [+ATI_Linux_Guys]: while Matthew is typing the answer are there any other general CATALYST questions Jon and I can field in the meantime?
[22:17:50] [@[R3D]MrB]: Why does CCC take so much memory lol sorry
[22:18:12] [+ATI_Linux_Guys]: Because we had a memory leak problem that will be fixed in the next CATALYST posting
[22:18:13] [@[R3D]Ratchet]: hehe you asked for it
[22:18:29] [+ATI_Linux_Guys]: its fine I dont mind answering generic stuff till Matthew answers the real question
[22:18:48] [@[r3d]ben6]: terry when can we expect CCC support for integrated chipsets like RS480?
[22:19:12] [+ATI_Linux_Guys]: If you search the Rage3D or Driverheaven forums you will see I have answered that already
[22:19:22] [+ATI_Linux_Guys]: But since Jon is watching me now I cant really say
[22:19:30] [+ATI_Linux_Guys]: J*A*N* cough cough
[22:19:50] [@[R3D]Ratchet]: what cool new features are coming up in CCC that you can tell us about
[22:19:57] [@[R3D]MrB]: that was a typo for Jon Terry ??
[22:20:20] [@[R3D]MrB]: What is the number of developers working over linux drivers? (terry might know? )
[22:21:03] [+ATI_Linux_Guys]: I do know but for competitive reasons we dont give direct numbers
[22:21:16] [+ATI_Linux_Guys]: All I can say is its roughly proportional to the size of the Linux market.
[22:21:27] [@[R3D]Ratchet]: would they all fit in a breadbox?
[22:21:30] [+mtippett]: stop
[22:21:44] [@[R3D]MrB]: your answer didn't get through matt
[22:22:08] [@[R3D]Ratchet]: you may need to break it up if it's too long
[22:22:13] [+mtippett]: Our primary focus is commercial GL customers. Enthusiasts and gamers can expect high value fixes such as the doom3 fix as they are needed. Keep in mind that a lot of the commercial customers are focused on performance as well and that will be available immeadiately for the enthusiasts.
[22:22:18] [+mtippett]: A benefit of OpenGL as the 3D subsystem for Linux is that the Workstation development becomes available immediately for the enthusiast gamer. Under Windows, the work involves OpenGL for workstation, DirectX for Gamers (as a general rule - ignoring doom3).
[22:22:19] [+mtippett]: stop
[22:22:20] [+mtippett]: :)
[22:22:33] [@[R3D]Ratchet]: <idr> QUESTION: On the control pannel subject, the open-source drivers have a standard XML based interface for exposing tunable options. Will ATI support that interface in their drivers?
[22:23:13] [+mtippett]: As indicated before, there are some technical pre-requisites, one of these would be considering the DRI configuration infrastructure (which is XML).
[22:23:36] [+mtippett]: No promises that it will be that though.
[22:23:45] [+mtippett]: URL -> http://dri.sourceforge.net/cgi-bin/moin.cgi/ConfigurationInfrastructure
[22:23:55] [+mtippett]: We haven't finished our analysis on what is needed.
[22:23:57] [+mtippett]: stop
[22:24:13] [@[R3D]Ratchet]: <Kaloz> QUESTION: AMD64 users are generally forced to use cards from a competitor to have decent 3D support on 64-bit Linux systems. Are there any (at least) beta drivers for AMD64 Linux, and if so, are there plans on making them public as the Windows beta drivers?
[22:24:57] ::: Mode: ([R3D]Ratchet) sets (+b *!*@
[22:24:57] ::: This (*!*@ ban affects (Kanniball)
[22:25:29] [+mtippett]: We are planning our first release of AMD64 with the next release.
[22:25:31] [+mtippett]: stop
[22:26:34] : +mtippett twiddles his thumbs.
[22:26:40] [@[R3D]Ratchet]: hehe
[22:26:55] : +mtippett explains /me to Terry
[22:26:58] : +ATI_Linux_Guys laughs at Matthew
[22:27:11] : +mtippett prods Ratchet...
[22:27:13] [@[R3D]Ratchet]: <HWolf> Question: Given ATI's business focus, The priority seems to be on 2d performance and functionality. Does ATI have any plans to support new desktop developments such as the X.org Composite extention?
[22:27:25] : +ATI_Linux_Guys are proud to have discovered "me" functions
[22:27:40] [+mtippett]: The Chats with Terry will never be the same...
[22:28:00] [+mtippett]: Don't quite understand the 2D performance implication.... But anyway...
[22:28:33] [+ATI_Linux_Guys]: just a little side note while Matthew answers for our "other OS" users
[22:28:46] [+ATI_Linux_Guys]: we are working on the punkbuster issue and I expect a hot fix posted
[22:28:50] [+ATI_Linux_Guys]: today, tomorrow or Monday
[22:29:21] [+mtippett]: As the technology becomes deployed in the market, we will respond to what is needed. Currently there is minimal requirements for Composite and so on, until some fundamental architectural changes are made.
[22:29:23] [+mtippett]: stop
[22:29:31] [@[R3D]Ratchet]: <rinthos> QUESTION: Back in Mid-August, 2004, at http://freedesktop.org/bugzilla/show_bug.cgi?id=1085#c4 , Ati stated that they would be re-releasing driver 3.12 with X.org 6.8 support in a few weeks. What is the status on this re-release?
[22:29:41] : +mtippett chuckles...
[22:29:48] [+mtippett]: I knew that would come up...
[22:30:01] : @[R3D]Ratchet knew too
[22:30:09] : +ATI_Linux_Guys didnt know
[22:30:18] : @[r3d]ben6 doesn't know anything
[22:31:19] [+mtippett]: That was a leak from beta program mailing list... It was 'planning' and not a definite statement. (I didn't say it, Stefan from SuSE did). We had other important things to work with that delayed that work.
[22:31:25] [+mtippett]: ATI did not state... Stefan did.
[22:31:48] [+mtippett]: Anyway, 6.8 is going to be in the next release.
[22:31:51] [+mtippett]: stop
[22:32:05] [@[R3D]Ratchet]: <adrianhensler> QUESTION: Is there any focus on implementing AIW features under linux?
[22:32:23] [+mtippett]: Currently there is no focus.
[22:32:23] [+mtippett]: stop
[22:32:33] [@[R3D]Ratchet]: <BSD_Diffie> QUESTION: Will ATi offer plain tar.gz/tar.bz2 package of the fglrx driver, if no why not?
[22:33:10] [+mtippett]: No. Because there are fundamental issues with providing tarballs.
[22:33:35] [+mtippett]: We are however working on our installer that will simplify things for a lot of people.
[22:34:56] [+mtippett]: In general, the distribution's packages are preferred for distributions that not directly supported by ATI. (SuSE/Debian/FC/Gentoo all have packages)
[22:34:59] [+mtippett]: stop
[22:35:07] [@[R3D]Ratchet]: QUESTION: Reading dissusions on the x.org mailing lists it seems X is heading in an OpenGL based direction with an Xgl implementation. Does ATI have any plans to support this?
[22:37:04] [+mtippett]: As said in the response to composite extension question. There are some fundamental technology issues that need to be solved before that becomes widely available. When Xgl becomes deployed and relevant, we will move in that direction.
[22:37:05] [+mtippett]: stop
[22:37:15] [@[R3D]Ratchet]: Any plans for BSD support?
[22:37:15] ::: Mode: (ChanServ) sets (+o [R3D]ben6)
[22:37:44] [+mtippett]: Ratchet: is that a question from you or the channel :)
[22:37:52] [@[R3D]Ratchet]: me, sort of
[22:37:59] [@[R3D]Ratchet]: so give a good answer ;)
[22:38:10] [+mtippett]: At this stage: no.
[22:38:22] [+mtippett]: stop
[22:38:46] [@[R3D]Ratchet]: <untamed> QUESTION: Does ATI plan to increase the resources it allows to the Linux developement or are they satisfied with the pace they have established so far?
[22:39:06] [+mtippett]: As market demand increases, the team will increase.
[22:39:32] [+mtippett]: In the last 12 months, demand _HAS_ increased, and hence the resource count has increased.
[22:39:36] [+mtippett]: stop
[22:39:54] [@[R3D]Ratchet]: <rinthos> QUESTION: Close to a year ago, Ati made an announcement they would be relocating their driver development to Canada as a means of increasing the productivity of their drivers, can you explain to us what benefits we, the end users, would notice as a result of this move?
[22:41:31] [+mtippett]: ATI _never_ made an official announcement of the move.
[22:42:02] [+mtippett]: But the driver development did move from Starnberg to Canada.
[22:42:32] [+mtippett]: The initial benefit is that I have a job
[22:42:35] : +mtippett smirks..
[22:42:37] [+mtippett]: Anyway.
[22:42:43] [@[R3D]Ratchet]: hehe
[22:42:46] [@[R3D]ben6]: heh
[22:42:46] [@[R3D]dementor]: lol
[22:42:49] : @[R3D]MrB ATI humour...
[22:43:02] [+mtippett]: The initial benefit is that the drivers are now released on a regular canter (bi-monthly).
[22:43:44] [+mtippett]: We use the same codebase at the time of release - CAT 4.11 windows drivers will have the same core code as the 8.08 Linux drivers
[22:44:02] [+mtippett]: We have direct access to HW and SW engineers, to resolve problems faster
[22:44:20] [+mtippett]: We have direct access to PR and Marketing.
[22:44:25] : +mtippett peers at Jon and Terry
[22:44:54] [@[R3D]ben6]: PR and marketing? Eww :)
[22:44:56] [+mtippett]: As you can see, the next drivers have a few extra goodies, which is a direct result of the work.
[22:44:59] [+mtippett]: stop
[22:45:23] [@[R3D]Ratchet]: QUESTION: How concerned is ATI about the effect the quality of it's linux drivers has on it's good name? This considering a lot of linux enthousiasts serve as a guide for less-technical friends and family in their purchasing decisions?
[22:45:41] [@[R3D]Ratchet]: And in this light, did ATI ever consider temporarily boosting the develment resources to bring the drivers up to par on 3d performance?
[22:45:42] [+mtippett]: (bi monthly = every two months - I am watching #ATILinuxQ.
[22:45:50] [@[R3D]MrB]: And in this light, did ATI ever consider temporarily boosting the develment resources to bring the drivers up to par on 3d performance?
[22:45:57] [+mtippett]: Three questions or one?
[22:46:06] [@[R3D]MrB]: oops my bad
[22:46:07] [@[R3D]Ratchet]: lets try again
[22:46:12] [@[R3D]Ratchet]: QUESTION: How concerned is ATI about the effect the quality of it's linux drivers has on it's good name? This considering a lot of linux enthousiasts serve as a guide for less-technical friends and family in their purchasing decisions? And in this light, did ATI ever consider temporarily boosting the develment resources to bring the drivers up to par on 3d performance?
[22:47:23] [+mtippett]: That is a PR question - Terry - Jon?
[22:47:55] [@[R3D]MrB]: well Jon comes up with PR spin Ratch do you want to ask another question?
[22:48:05] [+ATI_Linux_Guys]: You bet that we are both very concerned about anything that impacts the reputation of our software
[22:49:01] [+ATI_Linux_Guys]: as such we will now adopt a more open approach to keeping the user community update on our Linux developments, as we regularly update the press on our progress in the same manner we do with our windows drivers
[22:49:47] [+ATI_Linux_Guys]: In fact you will see more coverage on ATI's linux developments in the coming weeks at some of your favorite tech web sites and we will continue this trend moving forward
[22:49:48] [+ATI_Linux_Guys]: stop
[22:49:56] [+mtippett]: My response on the burst development. We are more interested in sustainable development than one-shot wonders, hence we want to ramp the team, but do it carefully.
[22:49:57] [@[R3D]Ratchet]: <agd5f> QUESTION: is there a chance flgrx will switch to using the open (xorg/sfree86) DDX (2d Driver) and continue with the closed 3d driver? THere are already quite a few bugs that still affect the flgrx driver that have been fixed in the xorg ddx
[22:49:57] [+mtippett]: stop
[22:50:02] [@[R3D]Ratchet]: <agd5f> QUESTION: is there a chance flgrx will switch to using the open (xorg/sfree86) DDX (2d Driver) and continue with the closed 3d driver? THere are already quite a few bugs that still affect the flgrx driver taht have been fixed in the xorg ddx
[22:50:40] [+mtippett]: No. There are proprietary extensions that we can not fold into the Xorg release.
[22:50:42] [+mtippett]: stop
[22:51:07] [@[R3D]Ratchet]: <Svartalf> QUESTION: PPC Support? PPC is one of the primary platforms- especially with IBM pushing it. What support is planned for PowerPC on the binary drivers?
[22:52:17] [+mtippett]: None at this stage.
[22:52:18] [+mtippett]: stop
[22:52:32] [@[R3D]Ratchet]: <Dazog> Question: Will ATI setup a Bugzilla for all of us in the Open source community wanting to help crush bugs and give you huge amount of information on fixing the Linux Drivers to overtake the competition?
[22:52:43] : @[R3D]Ratchet thinks that's one for Terry
[22:53:01] [+mtippett]: No. Our bug list will contain not only user problems, but also p
[22:53:02] [+mtippett]: stop
[22:53:17] [@[R3D]Ratchet]: <HWolf> QUESTION: Considering the fact the next release will be available for x86 aswell as x64, the amount of work to be done for each release will undoubtebly increase. will ATI employ additional resources for the x86x64 platform, or is this done with the same(sized) team?
[22:53:17] [+mtippett]: Erm..
[22:53:27] [+mtippett]: Let me retract.
[22:53:43] [@[R3D]Ratchet]: sure
[22:53:59] [+mtippett]: No. Our bug list will include issues other than end user problems. We will not seperate into two lists.
[22:54:01] [+mtippett]: stop
[22:54:03] [+mtippett]: :)
[22:54:06] [+mtippett]: Now the next question.
[22:54:19] [@[R3D]Ratchet]: <BSD_Diffie> QUESTION: As a Linux users who is in market for new video hardware, why should i choose ATi and purchase a video card from you over nVidia?
[22:54:29] [+mtippett]: Hold back Diffie's question...
[22:54:33] [+mtippett]: HWolf's question first...
[22:54:47] : +mtippett puts on the record
[22:54:59] [+mtippett]: As true customer demand increases, resourcing will increase
[22:55:03] : +mtippett turns off the record.
[22:55:06] [+mtippett]: stop
[22:55:09] [+mtippett]: Diffies question.
[22:55:18] : @[R3D]Ratchet is loosing control ;)
[22:55:33] [@[R3D]Ratchet]: only a few questions left
[22:55:37] [@[R3D]ben6]: We want to thank ati for joining us today. We have time for a couple more questions :). Thanks Terry and Jon and Matthew
[22:56:37] [+ATI_Linux_Guys]: Because as THE market leader in graphics, with award winning hardware across all product lines.
[22:56:49] [+ATI_Linux_Guys]: We have tackled our Windows driver problems of the past head on
[22:57:14] [+ATI_Linux_Guys]: and not only overcome those problems, we are now setting an unprecedented pace in Windows drivers
[22:57:36] [+ATI_Linux_Guys]: Innovations such as monthly releases, CCC, VPU recover, Smartgart, etc
[22:57:46] [@[R3D]ben6]: which reminds me. Isn't it about time to see 4.11 for Windows? :)
[22:57:50] [+ATI_Linux_Guys]: are all things we worked our a$$'S off to deliver
[22:58:00] [+ATI_Linux_Guys]: and we vow that the same passion will go into our Linux drivers
[22:58:05] [+ATI_Linux_Guys]: stop
[22:58:13] [@[R3D]Ratchet]: QUESTION: What are the issues in supporting ACPI sleep states? I understand that this is still under very heavy development but it seems that restoring video state is the main problem at the moment. Are the future drivers being overhauled to support ACPI sleep and suspend?
[22:59:03] [+mtippett]: We are working on our Power Management infrastructure, but as was mentioned, it is still evolving under Linux. I can't say when it will be in a truly supportable state.
[22:59:04] [+mtippett]: stop
[22:59:13] [+ATI_Linux_Guys]: Last question of the session
[22:59:19] [+ATI_Linux_Guys]: and then we have to sign off
[22:59:57] [+ATI_Linux_Guys]: oh and ben CAT 4.11 is scheduled sometime between now and a week from now
[23:00:00] [@[R3D]Ratchet]: ok, let me find a good one
[23:00:17] [@[R3D]Ratchet]: <WhO_KnOwS> QUESTION: When will running several X session be supported, if it will be supported at all?
[23:00:20] [@[R3D]ben6]: thanks terry :)
[23:00:44] : +mtippett smiles at Who_KnOwS
[23:01:32] [+mtippett]: It will be supported, but I don't have any firm dates.
[23:01:32] [+mtippett]: stop
[23:01:43] [@[R3D]Ratchet]: ok then, that's it I guess.
[23:01:47] [@[R3D]Ratchet]: thanks for your time guys
[23:01:49] [@[R3D]Ratchet]: :)
[23:01:56] [@[R3D]dementor]: Thanks for another solid chat
[23:01:56] [@[R3D]ben6]: thanks for coming
[23:02:02] [@[R3D]MrB]: it's not done
[23:02:07] [@[R3D]MrB]: matt has a few end words to say
[23:03:09] [+ATI_Linux_Guys]: I have a few words to say too
[23:03:14] [+mtippett]: If people really want to help us solve our problems. Work with each other to create a test case.
[23:04:13] [+mtippett]: If you can make the test case return 0 for return 1 for fail, then we can track it down easier.
[23:04:32] [+mtippett]: Post it through rage3d Linux/Drivers for the moment... Watch this space for more information in the coming months.
[23:04:33] [+mtippett]: stop
[23:05:33] [+ATI_Linux_Guys]: and from the Marketing/PR team, we want to thank Rage3D and all the mods, and you will be seeing more Linux related information in the coming weeks
[23:05:45] [+ATI_Linux_Guys]: I know for a fact Jon is lining up some interviews on some tech sites
[23:05:56] [+ATI_Linux_Guys]: so hopefully the information will get out to everyone
[23:06:02] [@[R3D]Ratchet]: and Rage3D would like to thank Terry, Jon, and especially Matt for doing this with us today.
[23:06:11] [+ATI_Linux_Guys]: and of course we all want to raise the quality of our Linux drivers to the next level
[23:06:18] [+ATI_Linux_Guys]: Cheers everyone see you next time
[23:06:24] [+mtippett]: Thanks guys
[23:06:34] [@[R3D]MrB]: and stop pounding rage3d
[23:06:37] [@[R3D]MrB]: what's up with that!!
[23:06:38] [@[R3D]MrB]: lol
[23:06:42] ::: Mode: ([R3D]MrB) sets (-m)