What to expect for 8.01 ?

[Nicolas]

So, what should we expect for the next release?

My guess is that at least the widescreen resolution problem should be fixed.

Hopefully AIGLX will be usable (scrolling for example).

What else?

[Kano]

Best expect nothing, then you will not be disappointed ;)

[Nicolas]

You're absolutelly right... But I've been expecting nothing for a few months (and didn't get disappointed, because nothing was delivered anyway).

Time to push it a little :p.

New year, maybe the devs are more motivated or something... ;)

[Uchikoma]

The some for everyone doesn't seem to apply for all people it seems haha.

I actually have no (showstopper) issues at present time, believe it or not. Improvements to the driver such that Compiz works a bit better I guess? That's about it really >_>

[yoshi314]

more bugs? that's guaranteed to succeed :]

[givemesugarr]

well, i expect that the resolution problem would be fixed and that there would be some other old known issues added to the known issues list. the last thing that i expect is that agp would finally be found and recognized by the driver.

[koolmanoncampus]

I am expecting widescreen to come back. Why 7.12 was even released with such a glaring usability problem is atrocious. I don't think anyone would have missed that one monthly release if it meant there was less time spent on tying everything up into a release and more on fixing problems.

And I'd hope at some level the developers spend time on important issues like proper video playback. I could give a damn about a wobbly window or a desktop cube when playing videos results in horrible playback ability(without AIGLX). My videos have horrible diagonal tearing. Using 7.11 now; which is nice because every openGL app leaks memory like crazy. Course 7.12 fixes it but the aforementioned resolution problem makes it not so worth it.

Ugh I just can't wait until there is support for xvideo and 3d on the open source side of things. Then I can began the phase in my life where I can forget about the horrible experiences I have had with fglrx.

[Bigon]

Hope that he resolution issue will be fixed and that compiz will work again with xorg 7.3

[rugger]

Ugh I just can't wait until there is support for xvideo and 3d on the open source side of things. Then I can began the phase in my life where I can forget about the horrible experiences I have had with fglrx.

Yep, maybe then I can pull the nvidia card I purchased (7300LE) and run my onboard ATI again (I really don't want high performance, just things to work properly)

[enzobelmont]

next month i expect some russian hackers finally build an usable Xpress 200M driver for MacOSX (hackint0sh)...
http://forum.insanelymac.com/index.php?showtopic=47944&st=160&gopid=569404&

eerr... sorry from _ATI_ i dont expect nothing but bugs.

some enthusiast groups put a lot more effort on building better graphic solutions than _ATI_

shame on you ati

my laptop has an integrated _ATI_, i can do nothing but be at _ati's_ will

sorry my fracking english...

[maggot_brain]

I've switched to RadeonHD, but if the next fglrx version fixes the switching VT bug and video tearing I might go back...until RadeonHD is fully functional of course.

[timmydog]

I've also switched to the radeonhd driver and it seems almost as good as the fglrx driver for video on my 690G and doesn't have the diagonal tear.

I used 8.43 for a while, but I can't cope without widescreen (my system is a PVR) and 7.11 didn't seem to have enough additional benefits for me to bother installing it. :-(

Here's hoping that xv works on amd64 systems!

[givemesugarr]

I've also switched to the radeonhd driver and it seems almost as good as the fglrx driver for video on my 690G and doesn't have the diagonal tear.

I used 8.43 for a while, but I can't cope without widescreen (my system is a PVR) and 7.11 didn't seem to have enough additional benefits for me to bother installing it. :-(

Here's hoping that xv works on amd64 systems!

well, there was a report of someone getting the 1650x1080 to work after forcing the prefered mode via a specified modeline in the modes section. try looking at the last pages of the 7.12 thread and you could try out the howto about extracting the modeline from your xorg.log and try use it as indicated there. maybe you'll be lucky and have it work.
xv seems to work but only for some boards but it's slow when it does.

[Compxpert]

I'm sorry but exactly what do you all need XV for. I use opengl for video. Is XV better or something.

On topic: I'd like native 1680x1050 support and not have to use a modeline to get it. I'd also like to see AIGLX working properly and compiz running properly. (Perhaps these problems go hand in hand?) As its ATI though I can't expect too much from them. However I suppose if you go into it not expecting much of anything and you end up getting a lot its a nice surprise but again don't expect much.

[Xemanth]

I would like to see X1100 graphics work in 2d without graphical artifacts ^^
But I'm almost sure that there will be no fix

[bridgman]

Are you talking about corruption around the cursor and in the lower right hand corner, or some other kind of artifacts ?

[cjr2k3]

Sometimes I get the corruption around the cursor. Nothing special, normaly disapear after changing window or so.

[givemesugarr]

Sometimes I get the corruption around the cursor. Nothing special, normaly disapear after changing window or so.

i think this is due to some refresh rate and offscreen issue. after adding the xaanooffscreen option in xorg.conf you sometime experience this with boards that seem to have low values of fps.

[kelnoky]

I sincerly hope that they fix the really annoying gamma bug, which was only partially fixed, at least for me.
I expect nothing and I can say that if the next one or two drivers will be just as bad as the others before I am going to buy an nVidia card.

[tsuru]

I'm running on a lowly Mobility 9600 (M10). I was happy when the new speed boost came with the rewrite however the latest 7-12 has killed performance on 3D apps, namely Wings3D. Further more when I program in OpenGL, accessing framebuffer objects, EXT_framebuffer_object, and VBOs also is generating errors with code known to work in the past.

Having the issues affecting OpenGL programming is what I'd like to see fixed.

Btw. I just bought a 22 inch 1680x1050 native rez monitor and it will display widescreen. So did my 1280x800 lvdms. However I will concede there appears to be a problem with EDID handling. My lvdms EDID is either undetectable or non-existent so Catalyst never showed a problem. My monitor however does have detectable EDID and I had to add an Option ForceMonitors (nolvdms, crt1) to my xorg.conf for flgrx to work.

[Svartalf]

I'm running on a lowly Mobility 9600 (M10). I was happy when the new speed boost came with the rewrite however the latest 7-12 has killed performance on 3D apps, namely Wings3D. Further more when I program in OpenGL, accessing framebuffer objects, EXT_framebuffer_object, and VBOs also is generating errors with code known to work in the past.


Heh... The VBO problems are going to be obnoxious. The FBO stuff may be related to the fact that the Vista/XP drivers didn't have pretty much any of that properly implemented when they shipped (and I think it's still not 100% there on those platforms...).

[Xemanth]

Are you talking about corruption around the cursor and in the lower right hand corner, or some other kind of artifacts ?

Around cursor mainly but there's weird corruption also in window corners. I'll do couple screenshots later today.

[domi]

I'm sorry but exactly what do you all need XV for. I use opengl for video. Is XV better or something.

With 7.12, (radeon X1650, Xorg 7.2), opengl performance are quite bad (100% CPU for a 720p/x264 clip). With XV (and tearing :( ), CPU usage is about 30%

Note that opengl perf were much better with 8.42.

HTH

[Karel]

Best expect nothing, then you will not be disappointed ;)


Not really - expecting nothing is quite optimistic, they will introduce new bugs just so that they can say "We added new features".

[timmydog]

well, there was a report of someone getting the 1650x1080 to work after forcing the prefered mode via a specified modeline in the modes section. try looking at the last pages of the 7.12 thread and you could try out the howto about extracting the modeline from your xorg.log and try use it as indicated there. maybe you'll be lucky and have it work.
xv seems to work but only for some boards but it's slow when it does.

Thanks for the pointer - I read that thread too. ;-)

I've been hacking about with x config files for years and am just getting used to the fact that I don't need to (very much) anymore.

Since the bug is already known, I thought that I'd hold off until 8.01, continue using RadeonHD and spend more time with my wife during the holidays.

BTW I really do mean thanks for the pointer - you've been a star helping people out on the list!

Tim

[greg]

I'm sorry but exactly what do you all need XV for. I use opengl for video. Is XV better or something.

Xv is the standard interface for accelerated video scaling. Many applications depend on it, there's no excuse for not having it nowadays. It's faster than OpenGL in most cases, too.

[givemesugarr]

Xv is the standard interface for accelerated video scaling. Many applications depend on it, there's no excuse for not having it nowadays. It's faster than OpenGL in most cases, too.

not in the fglrx one... :rolleyes:

[bridgman]

It should be faster, other than on an RS690 in 64-bit or on a pre-5xx workstation part where you need to use the overlay for OpenGL workstation apps.

Doesn't hurt to check settings :

- for 5xx and above cores (including RS690) you want "Video Overlay" off and "Textured Video" on -- I don't think the "OpenGL Overlay" setting matters but not 100% sure

If you are having DRM/OpenGL config problems fix those first, since you won't get video acceleration unless the 3d stack is working properly

- for pre-5xx cores (including RS480) you want "Textured Video" off, Video Overlay" on, and "OpenGL Overlay" off.

Pre-5xx video is, afaik, independent of the DRM/3d stack.

[givemesugarr]

It should be faster, other than on an RS690 in 64-bit or on a pre-5xx workstation part where you need to use the overlay for OpenGL workstation apps.

Doesn't hurt to check settings :

- for 5xx and above cores (including RS690) you want "Video Overlay" off and "Textured Video" on -- I don't think the "OpenGL Overlay" setting matters but not 100% sure

If you are having DRM/OpenGL config problems fix those first, since you won't get video acceleration unless the 3d stack is working properly

- for pre-5xx cores (including RS480) you want "Textured Video" off, Video Overlay" on, and "OpenGL Overlay" off.

Pre-5xx video is, afaik, independent of the DRM/3d stack.

my experience on x200m is that opengl 2d-tex-fragprog works faster but with the lines onscreen issue, while 2d-tex works faster but the text is rendered in a very bad way. xv on the contrary works but it's quite slow when compared to opengl. the textured video to me seems to work better set to on in conjuction with texturedxrender on. on the hd2600 instead xv seems to be quite stable and works fine. the opengl has the same issues as on the rs480 (so they're in the driver somewhere) but is faster than xv.
on the other board i never tried the 3d acceleration but the overlay is set to xv and it works quite smooth and without any major issues.
now, after confronting what i've experienced and what you've proposed as configs it seems that something is not right either in your configs or in mine... or maybe in the driver itself. :confused:

[bridgman]

I think we're saying the same thing, except you're seeing that on the 2600 "opengl is faster but not as good quality" ? That may be correct -- what I was trying to say was "textured video should be faster for the same amount of video processing / quality" or something like that, since that's where we put the video-specific code.

On the X200M (RS4xx) I think the recommended settings would be "Textured everything" off and "Video Overlay" on. Textured Video really became a focus with the 5xx+ parts AFAIK. Are you saying that "opengl 2d-tex-fragprog works faster" on x200m than xv with the overlay, or just faster than xv with Textured Video ?

[givemesugarr]

I think we're saying the same thing, except you're seeing that on the 2600 "opengl is faster but not as good quality" ? That may be correct -- what I was trying to say was "textured video should be faster for the same amount of video processing / quality" or something like that, since that's where we put the video-specific code.

yes. the quality of opengl compared to xv is lesser for the lining issues and for the text rendering, especially with osd text. this happens with both 2600 and rs480. but in the games that work on opengl the 2600 is quite good and doesn't seem to present issues, except for the gamma correction. also i get some minor problems with the compiz cube rendering (sometime i get slow response and some watermark) when using the bigdesktop. these are the problems that i encounter with the 2600 on which i'm using 7.11 since i wasn't experiencing the opengl leak for some reason. or better, i'd experienced it for the first 30 seconds then it wouldn't continue to leak. i don't know why.


On the X200M (RS4xx) I think the recommended settings would be "Textured everything" off and "Video Overlay" on. Textured Video really became a focus with the 5xx+ parts AFAIK. Are you saying that "opengl 2d-tex-fragprog works faster" on x200m than xv with the overlay, or just faster than xv with Textured Video ?

i normally use opengl only in xine on the x200 since that's the only way of having something useful. the 2d-tex-fragprog is faster than xv allowing me to play also hd format video content (i've tried some 1920x1080 h264 and they played as in windows). this is impossibile with the xv. i also seen that the overall processor load on opengl is lower than on xv. but this is just when using the fragprog on the text rendering. the problem is that fglrx doesn't support fragprog text rendering. it doesn't work on text not embeded directly into the video-stream, and this makes me suspect the driver doesn't support it. the 2d-tex rendering instead renders well the text but doesn't have good performance. it seems like the first mode just ignores the osds while the second one doesn't have the speed. the strange thing is that if i try to enable textured2d then xorg would crash. it should work too if i get the rendering options enabled. :confused:

well, it seems that we were both saying about the same thing on the configs in the end, and my hunger made me understand it in the wrong way. :)

by the way is there a chance to get the new fglrx next week or is more likely to come out the following one considering christmas and new years eve?

[SirYes]

the last thing that i expect is that agp would finally be found and recognized by the driver.

Amen, brother!

[Tsume]

Just to clarify...

Widescreen is not broken, but the 1650x1080 resolution is broken.

My dual setup of 1440x900 monitors works fine; as far as widescreen goes anyway. All of the other bugs (video tearing, crappy AIGLX, etc.) are there.

[bridgman]

i normally use opengl only in xine on the x200 since that's the only way of having something useful. the 2d-tex-fragprog is faster than xv allowing me to play also hd format video content (i've tried some 1920x1080 h264 and they played as in windows). this is impossibile with the xv.

Good point. "Faster and not so good" is useful at higher resolutions, even if it's... um... not so good.

[givemesugarr]

Good point. "Faster and not so good" is useful at higher resolutions, even if it's... um... not so good.

maybe i had to say that the hd content was playable but not watchable.... the difference of faster was in the frames dropped: about 60-65% on xv and about 50% on opengl. this is what means faster on the x200m. and it works in the same way on windows, so this i not a driver problem.
the fact of opengl is clear on the hd2600: its slower and the processor is very stressed when using hd content on linux when compared to windows. the content this time is watchable but the processor need to be at least a mid range dualcore or you risk not having the average needed quality. thanks god that hd content is quite rare and that i've tried it just to see how it works. so the faster is always true, the not so good is to be taken as it is: not playable/not watchable on rs480 playable/watchable on hd2600. on the other chip i haven't tested it yet.

[peol]

Just to clarify...

Widescreen is not broken, but the 1650x1080 resolution is broken.

My dual setup of 1440x900 monitors works fine; as far as widescreen goes anyway. All of the other bugs (video tearing, crappy AIGLX, etc.) are there.

Well, widescreen sure ain't working over here on my flatscreen (1440x900 is native). I'm stuck at 1280x1024.

I've been patient enough and now when the new 8800GT arrived (the Call of Duty-bundled one), there's no reason for me to sit and wait for something to happen.

[Tsume]

Well, widescreen sure ain't working over here on my flatscreen (1440x900 is native). I'm stuck at 1280x1024.

I've been patient enough and now when the new 8800GT arrived (the Call of Duty-bundled one), there's no reason for me to sit and wait for something to happen.

Weird...

1440x900 on both my screens is working fine. Are you using DVI or VGA? I'm doing dual VGA.

[DanL]

I predict people will continue to b1tch and threaten to buy nvidia regardless of how good the drivers are. Personally, I'm using radeonhd and I've left fglrx forever (and haven't looked back). BTW, as I said in the X1250 Google group, I've seen tons of issues in the Ubuntu forums about nvidia. The grass isn't all green on that side of the fence either.

[Kano]

The Ubuntu problems with gfx drivers are house made, because of the way Ubuntu uses restricted modules. That's a too complex way for beginners who try to switch between drivers.

[wiscados]

I expect all the software on the standard Ubuntu installation to be Free.

[givemesugarr]

I expect all the software on the standard Ubuntu installation to be Free.

fglrx is free.... you don't have to pay to have it....

[bridgman]

Did someone mention free beer ?

http://www.curvesforum.com/forum/images/smilies/new05b/popcorn.gif

[Svartalf]

Did someone mention free beer ?

http://www.curvesforum.com/forum/images/smilies/new05b/popcorn.gif

Heh... Where's the party starting, then? >:-) :D

If it's all going to be free, pour me a tall, tall glass of Arrogant Bastard Ale...

[DanL]

If it's all going to be free, pour me a tall, tall glass of Arrogant Bastard Ale...

I prefer a nice, dark Pretentious Asshole Lager.

[Svartalf]

I prefer a nice, dark Pretentious Asshole Lager.

Riight... You think you're being funny...

Arrogant Bastard Ale (http://www.arrogantbastard.com) is actually a brand name for one of the better microbrews in the US. I was actually being honest when I said what I said.

[DanL]

Riight... You think you're being funny...

Believe it or not, I realized yours was an actual beer and mine was not. I was doing a bit of social commentary on the various types of jerks one might find on the interweb. I guess I failed. :(

[Michael]

Arrogant Bastard Ale (http://www.arrogantbastard.com) is actually a brand name for one of the better microbrews in the US. I was actually being honest when I said what I said.


++ for Arrogant Bastard Ale, but let's not get this thread off topic with our favorite brews :)

[pedepy]

++ for Arrogant Bastard Ale, but let's not get this thread off topic with our favorite brews :)

it was bound to happen at some point, the day ATI linux users would resort to alcohol to forget about fglrx.. enough of these babies make aiglx appear smoother..

oh well

:(

(oh and uh ++ for dark pretentious brews) :)

[Svartalf]

++ for Arrogant Bastard Ale, but let's not get this thread off topic with our favorite brews :)

Not a problem Michael. I was just chiming in with a bit of fun banter and ended up having to nail someone's foot to the floor for "making social commentary" at my expense.

To DanL: You might want to try your hand at something else. You just simply ended up making yourself look like one of those jerks you were "commenting" on.

[Svartalf]

And to get this back on topic:

I expect them to fix a few minor things because of a distinct lack of sufficient resources to handle going through and auditing the code based off of Valgrind reports and Oprofile tuning reports- which is what they're needing to do.

What I hope they'll do is fix the Rialto and the X200M issues. What will happen remains to be seen.

[Svartalf]

it was bound to happen at some point, the day ATI linux users would resort to alcohol to forget about fglrx.. enough of these babies make aiglx appear smoother..

< *snicker* > :D

[Draku]

Catalyst 8.1 for Windows Released Now http://game.amd.com/us-en/drivers_catalyst.aspx

Still waiting for linux version : http://ati.amd.com/support/drivers/linux/linux-radeon.html

[Svartalf]

Catalyst 8.1 for Windows Released Now http://game.amd.com/us-en/drivers_catalyst.aspx

Still waiting for linux version : http://ati.amd.com/support/drivers/linux/linux-radeon.html

It depends on how far along their SQA people are on testing and whatnot. I'm betting it'll be Friday at the earliest, more probably sometime next week before we see what comes out of AMD. I'm hoping to have a handful of things fixed- we'll see. I would like to be able to verify function on several titles for Michael Simms with the proprietary drivers. With all the ongoing issues from prior iterations, I've kind of held off a bit because it'd been a waste of my time.

[Draku]

After updating my pc i don't have any AGP problems :)

[ddoan]

I'm not sure if it is known... but one can find some interesting information in the release notes of the Windows version... have a look at the last sentence:

"Catalyst 8.1 introduces MultiView support. This feature provides for hardware accelerated OpenGL rendering across multiple graphics adapters. MultiView will provide hardware accelerated 3D rendering in a system containing multiple graphics cards on an extended desktop arrangement. This feature will allow for the rendering performance and additional frame buffer resources to be evenly shared with the second and third graphics adapters. This allows for a 3D application to have the same performance running on a secondary or third display device as if it were running on the primary display device. This feature will be supported under Windows XP (32 and 64 bit versions), Windows Vista (32 and 64 bit versions), and the Linux operating system."

Sorry, if this is already known ... ;-)

[Svartalf]

"Catalyst 8.1 introduces MultiView support. This feature provides for hardware accelerated OpenGL rendering across multiple graphics adapters. MultiView will provide hardware accelerated 3D rendering in a system containing multiple graphics cards on an extended desktop arrangement. This feature will allow for the rendering performance and additional frame buffer resources to be evenly shared with the second and third graphics adapters. This allows for a 3D application to have the same performance running on a secondary or third display device as if it were running on the primary display device. This feature will be supported under Windows XP (32 and 64 bit versions), Windows Vista (32 and 64 bit versions), and the Linux operating system."

Interesting. Very interesting. I would hesitate to claim the same performance unless the GPU's are the same on the two or three cards, but it's a nice feature if you're not needing something like Crossfire or SLI.

[BlackStar]

"On the Linux side of things, AMD says it plans to release Catalyst 8.1 drivers for Linux soon. That release will include MultiView, too, and it will introduce support for Ubuntu 7.10 as well as versions 5.1 and 4.6 of Red Hat's Enterprise Linux. In addition, the upcoming drivers will resolve problems related to display corruption after periods of inactivity, support for displays with 1680 x 1050 resolutions, ignored lines in the xorg.conf configuration file, and setting a system to sleep or hibernate. Expect the Linux drivers to be available from AMD's drivers and software page."

Source (http://www.techreport.com/discussions.x/13946)

[fhuberts]

http://www.phoronix.com/forums/showthread.php?t=7416

[Svartalf]

http://www.phoronix.com/forums/showthread.php?t=7416

Perhaps they will propagate the AGP fixes they recently figured out from the Windows side to the Linux side.

[Uchikoma]

I see sleep and hibernate. If it happens, I'm happy for another month ^^;

[MagicTempest]

I never use sleep or hibernate on my laptop, but I might begin using it if ATI fixes those problems. But I'd rather see the AGP fixes and a fix for the 1680x1050 problem, but according to your source those problems will all be solved

[Michael]

The 1680x1050 problem fix / non-VESA mode problem is FIXED in 8.01.

[lucky_]

8.01 wiil be available for firegl ?
I ask because last year firegls were really supported in many few releases.

[Michael]

8.01 wiil be available for firegl ?
I ask because last year firegls were really supported in many few releases.

Yes, it will. The lack of FireGL support came down to the implementation of the new fglrx driver -- so future releases should all support the FireGL ASICs but as has been the case, there is about a release every quarter that is "certified" for the workstation usage.

[lucky_]

Yes, it will. The lack of FireGL support came down to the implementation of the new fglrx driver -- so future releases should all support the FireGL ASICs but as has been the case, there is about a release every quarter that is "certified" for the workstation usage.
Cool, the certified 7.12 scrambled the screen on a fresh install of ubuntu.
I could not read a single line once the desktop is loaded. (on gdm it's fine, and I don't think to have compiz active )

[GuiPoM]

Yes, it will. The lack of FireGL support came down to the implementation of the new fglrx driver -- so future releases should all support the FireGL ASICs but as has been the case, there is about a release every quarter that is "certified" for the workstation usage.

Has anyone be able to activate composite on a FireGL with 7.12 drivers ? I guess it will not be fixed on 8.01, but opengl overlay is forced to on when a firegl board is detected, what prevents composite to be used.

Any idea if it is a knwon bug/fix ?

[Draku]

Linux 8.1 drivers online :http://ati.amd.com/support/drivers/linux/linux-radeon.html

New Supported Operating Systems

This release of the ATI Catalyst Linux software suite introduces support for the following new operating systems

* RedHat Enteprise Linux 5.1
* RedHat Enteprise Linux 4.6
* Ubuntu 7.10 (Gutsy Gibbon)

Resolved Issues

The following section provide a brief description of resolved issues with the latest version of the ATI Catalyst™ Linux software suite. These include:

* Corruption will no longer be noticed in the lower right corner of the display or on the mouse pointer after the system is running for a long period of time
* Connecting a display device that supports 1680x1050 to a system running Linux will no longer result in a maximum display resolution of 1280x1024 only being available
* Custom mode lines in xorg.conf will no longer be ignored by the fglrx driver
* Suspending to RAM or DISK on kernel version 2.6.23 or later no longer fails

Known Issues

The following section provides a brief description of known issues associated with the latest version of ATI Catalyst™ Linux software suite. These issues include:

* On workstation hardware 3D applications will be corrupted if the screen width is not an integer multiple of 64 pixels, for example with a 1680x1050 wide screen display. Further details can be found in topic number 737-31720
* There is no support for video playback on the second head in dual head mode. Further details can be found in topic number 737-26985
* Desktop corruption may be noticed when dragging the overlay/video when using dual-display mode. Further details can be found in topic number 737-29578
* A black screen may be observed on some hardware when switching to the console or leaving the X window system when a Vesa framebuffer console driver is used. Further details can be found in topic number 737-30687
* Display flicker may be noticed when the gnome screen-saver starts
* Diagonal tearing may be noticed when playing a video file using a video player that utilizes the XVideo extension
* Video playback may look blocky when playing a video file using a video player that utilizes the XVideo extension
* Video Playback may display wrong colors and additional shadow images when cropping or expanding a video file using a video player that utilizes the XVideo extension
* Building RPM packages for Mandriva may fail

[Bigon]

I should have asked 3d acceleration *grrr*