Any news on ATI fglrx 8.42? [Archive] - Page 3

Snake

10-19-2007, 01:57 AM

Hi everyone.

I shouldn't talk about it but I'll say it anyway : We (at ATI) are going to provide the best fglrx drivers ever for Linux with EVERYTHING fixed and AIGLX perfectly working.

I'll probably get fired after telling you this because of my NDA with ATI, sorry, AMD, but here it is : get prepared for these amazing new drivers !

+++ AMD/ATI fired sole linux driver developer +++

After discovering a non-disclosure violation on the Phoronix forums commited by one of their developers, AMD/ATI fired the developer of their linux graphics driver software (aka "fglrx"). The release of the already announced 8.42 driver version will be delayed until a new developer is hired and has build up a deep understanding of the currrent codebase.

That's what you get for pressing people, folks! ;)

(For those who did not get it: THIS IS A JOKE (hopefully)).

Karel

10-19-2007, 03:14 AM

+++ AMD/ATI fired sole linux driver developer +++

After discovering a non-disclosure violation on the Phoronix forums commited by one of their developers, AMD/ATI fired the developer of their linux graphics driver software (aka "fglrx"). The release of the already announced 8.42 driver version will be delayed until a new developer is hired and has build up a deep understanding of the currrent codebase.

That's what you get for pressing people, folks! ;)

(For those who did not get it: THIS IS A JOKE (hopefully)).

The sole developer or the NDA is a joke? :)

fredo

10-19-2007, 04:14 AM

Unfortunately for the Linux community, this is not a joke. I've just been fired by AMD a couple hours ago.

The bad thing is that I didn't have the time to release the 8.42 driver so it probably won't be released this month or even this year...

slickhenry

10-19-2007, 04:35 AM

(10 + 21 + 28 + 18 + 31 + 26 + 19 + 14 + 12) / 9 = approx. 19.888889

Statistically it will be released late today.

But don't take this as gospel

givemesugarr

10-19-2007, 04:47 AM

Statistically it will be released late today.

the problem is how late..... :confused:

slickhenry

10-19-2007, 05:07 AM

the problem is how late..... :confused:

(ans - 19) * 24 = approx. 21.333333

So it should be released oct 19th at 21.33 (9.33pm). (doesn't say what timezone, tho')

:rolleyes:

sundown

10-19-2007, 06:01 AM

Karel

10-19-2007, 07:11 AM

Yes, they just fire you in one day nowdays :eek:

I think you can - at least if it's McDonalds type of job... I leave it to your judgement whether fglrx has more quality (and requirements) than a result of high school afternoon Pascal programming class.

EnderWiggin

10-19-2007, 09:20 AM

An infinite number of monkeys with infinite available time will produce working driver too... I wonder whether ATI/AMD or the monkeys will be faster.

The good news is that if you have an infinite number of monkeys, they will produce the driver instantaneously. Unfortunately ATI only has a few monkeys and is going for the latter of the truths.

Karel

10-19-2007, 10:49 AM

The good news is that if you have an infinite number of monkeys, they will produce the driver instantaneously. Unfortunately ATI only has a few monkeys and is going for the latter of the truths.

I'm not sure about that - unless you count the probability that a monkey will code the driver in one single point of time. However, my calculation is different - I have an infinite number of monkeys repeatedly bashing the keyboard (as you cannot produce whole driver in one moment) and I believe that if we give them a certain amount of time (infinite), one monkey will eventually be right (OK, an infinite number of monkeys will have infinite number of correct drivers as the number of possibilities to code a driver is not finite).

Well, the important thing is ATI has a finite number of monkeys and (as they think) infinite amount of time... I am not sure if that is going to work...

Uchikoma

10-19-2007, 10:53 AM

I hereby declare that the driver will not come out today.

I have...lets see...5-8 hours gap in which they might put it up if they leave office at 4ish. So add a couple hours (7-9 hours) from time of posting and we'll call it next week.

(Infinite) monkeys may get the job done, but is it a job WELL done? :)

Extreme Coder

10-19-2007, 11:38 AM

I sure am glad fredo was joking (he is, right? :P ) or else I'd be buying an nVidia card in 8.42 minutes exactly ;)

Extreme Coder

10-19-2007, 11:41 AM

I'm just asking for one thing in this driver:
better peformance(even than in 8.41)
When I tried 8.41, i was pleased that performance increased(glxgears went from 800-900 to 1100-1200)
but still some games are slower than they are in Windows. (same game, exactly same settings)
So Michael tell me, is there a possibility of better performance in 8.42?

Next on my list would be AIGLX ;)

fredo

10-19-2007, 11:43 AM

I sure am glad fredo was joking (he is, right? :P ) or else I'd be buying an nVidia card in 8.42 minutes exactly ;)

Well, I was indeed joking.

I haven't been fired so get prepared for THE BEST DRIVER EVER :D

JackDanielz

10-19-2007, 11:47 AM

Well, I was indeed joking.

I haven't been fired so get prepared for THE BEST DRIVER EVER :D

Wow that would be nice if anyone really believed you fredo.

EnderWiggin

10-19-2007, 11:50 AM

(Infinite) monkeys may get the job done, but is it a job WELL done? :)

Theoretically there would be a perfect driver hidden among the infinitely bad ones. Okay, actually there would be infinitely many perfect drivers, but we would never be able to find them because there are infinitely many drivers in general. Perhaps we need an infinite improbability drive to sort things out??

What if the monkeys are allowed to breed for several generations and the selection criteria is based on coding ability? How do genetic algorithms cope with infinity?

Hope these problems can keep us at bay until Monday. ;)

JackDanielz

10-19-2007, 11:53 AM

I wish I hadn't found out about this until it was already released.

Karel

10-19-2007, 11:54 AM

Theoretically there would be a perfect driver hidden among the infinitely bad ones. Okay, actually there would be infinitely many perfect drivers, but we would never be able to find them because there are infinitely many drivers in general. Perhaps we need an infinite improbability drive to sort things out??

What if the monkeys are allowed to breed for several generations and the selection criteria is based on coding ability? How do genetic algorithms cope with infinity?

Hope these problems can keep us at bay until Monday. ;)

No, they can't. Look, brain of a size of a planet, yet I can't use my ATI card at full speed. I feel so depressed. Let's hope, the first fifty million years of waiting will be the worst and from that time on, everything will haven been good ;)

I'm not sure why the driver is delayed, there was already a Phoronix article ready and written (now hidden only in Digg cache, if it isn't a fake).

Nicolas

10-19-2007, 11:56 AM

Just remembered something that could be improved in fglrx: PowerPlay support... My power consumption in Windows is WAY lower than Linux, and I think this is mostly ATI's fault.

MagicTempest

10-19-2007, 12:03 PM

Where has it been said that the driver is delayed? Ati has a release schedule in which they release a driver roughly every month. Nowhere has there ever been a date stated on which the driver should be released. Therefore they're not late or delayed.

You might be able to say they're late when they don't release the driver before the last day of this month, but as far as I know the release schedule is not fixed so they could skip a month when necessary.

Michael

10-19-2007, 12:35 PM

To address the concerns of users, the 8.42 driver will be released soon and contrary to rumors is not still in "heavy development" nor was it delayed because of "severe" issues. Any rumors about a leaked article by Phoronix is incorrect, but an outside party had illegally acquired content. As always, a Phoronix review of the new driver will appear on the day that the driver is being made available to the public.

Uchikoma

10-19-2007, 12:40 PM

Just remembered something that could be improved in fglrx: PowerPlay support... My power consumption in Windows is WAY lower than Linux, and I think this is mostly ATI's fault.

Well hmm. There is PowerPlay support. To view you're current PowerPlay setting:
aticonfig --list-powerstate

To set your powerstate to something
aticonfig --set-powerstate=1
...replace 1 with the desired selection. For the list of selections refer to the above list-powerstate command.

In the off chance that you know this, then I dunno what card you have, but mine works perfectly good (and I can get upto 6 hours of battery life on my laptop)

...back to the (infinite) monkey problem. There are how many lines of code in a driver? Okay...so say...oh I dunno. Some tens of thousands of lines (because I really don't know numbers with regards to that). Now, given that there are just that many individual letters, what are the chances that a monkey will be able to batter out each correct individual letter making that perfect driver?

givemesugarr

10-19-2007, 12:46 PM

Wow that would be nice if anyone really believed you fredo.

i really did believed fredo..... :p

Theoretically there would be a perfect driver hidden among the infinitely bad ones. Okay, actually there would be infinitely many perfect drivers, but we would never be able to find them because there are infinitely many drivers in general. Perhaps we need an infinite improbability drive to sort things out??

What if the monkeys are allowed to breed for several generations and the selection criteria is based on coding ability? How do genetic algorithms cope with infinity?

Hope these problems can keep us at bay until Monday.

the new ati driver will contain a very astonishing feature:
the answer to the fundamental question... :D:D:D:D

obviously, mankind is not fully prepared to receive that info, so ati is delaying the release date to better prepare mankind to this unique event.

and obviously, now that they finally got the answer to the fundamental question, they've forgotten the question, so they've decided to release specs to an outer group to provide mankind with the real fundamental question. :D:D:D:D:

Nicolas

10-19-2007, 12:47 PM

Well hmm. There is PowerPlay support. To view you're current PowerPlay setting:
aticonfig --list-powerstate

To set your powerstate to something
aticonfig --set-powerstate=1
...replace 1 with the desired selection. For the list of selections refer to the above list-powerstate command.

In the off chance that you know this, then I dunno what card you have, but mine works perfectly good (and I can get upto 6 hours of battery life on my laptop)

...back to the (infinite) monkey problem. There are how many lines of code in a driver? Okay...so say...oh I dunno. Some tens of thousands of lines (because I really don't know numbers with regards to that). Now, given that there are just that many individual letters, what are the chances that a monkey will be able to batter out each correct individual letter making that perfect driver?

Yes, I know this, but if I'm not mistaken, fglrx's PowerPlay features are not as good as Windows. I believe Linux's implementation lowers clock speeds and voltages, while the Windows driver is also able to "shut down" some parts of the core.
And also, I don't want to have to change my card's speed manually. It should be done automatically.

I have a Mobility X700 128MB card on my notebook.

givemesugarr

10-19-2007, 01:00 PM

And also, I don't want to have to change my card's speed manually. It should be done automatically.

I have a Mobility X700 128MB card on my notebook.

it would be nice if some developer could do a qt interface to the possible options.
maybe when i get a little time i will start doing some testing with qtdesigner, but first i'd need to know what settings can be done for fglrx.
actually i think that i use less than 2/3 of the entire configuration settings that i would be able to actually use.
in fact the ati control center for linux is still quite limited in this, but i would prefer to have better drivers than have the functions added to a control pannel.
anyway, michael, when you'd review the new driver i'd like to know if compiz-fusion would be able to run at decent speed with old xpress200m cards. i have that board and it's a horrible one. the opensource driver doesn't support it well and rendering is very experimental, aiglx works but compiz or beryl does not.
also i had to install the svn xorg-server to get the opensource driver to work with aiglx, so it would be nice to test the compatibility of the new driver with the new 1.4 server and new xorg 7.3. i don't espect this version to work but it would be a nice surprise if it would.

kirù

10-19-2007, 01:02 PM

http://www.phoronix.com/scan.php?page=news_item&px=NjEzOA

lol, 1000000 nvidia drivers will be released before the ati one. :(:(:(

Uchikoma

10-19-2007, 01:19 PM

Yes, I know this, but if I'm not mistaken, fglrx's PowerPlay features are not as good as Windows. I believe Linux's implementation lowers clock speeds and voltages, while the Windows driver is also able to "shut down" some parts of the core.
And also, I don't want to have to change my card's speed manually. It should be done automatically.

I have a Mobility X700 128MB card on my notebook.

I just wrote a script that autochanges it to lowest setting on startup. I don't game too much yet, so it's suits my needs.

...as for battery life, I get roughly the equivalent battery life as Windows, so I don't see a difference.

Svartalf

10-19-2007, 01:27 PM

i really did believed fredo..... :p

I've got a bridge in New York City to sell ya. >:-) :D

Nobody in their right or not so right mind would be chiming in any more than we are (or, for that matter, I have...) on this subject.

Now, having said this... I'm sure that the driver will be out before end of month; as to whether it'll be a major improvement (Not that this would be hard for them- just unlock and support R300 on up would be that for them at this point...) remains to be seen.

The_Monkey_King

10-19-2007, 01:38 PM

Dambit! All my monkeys produce is "The Taming of the Shrew." They are no where near coding... :(
An infinite number of monkeys with infinite available time will produce working driver too... I wonder whether ATI/AMD or the monkeys will be faster.

givemesugarr

10-19-2007, 01:44 PM

I've got a bridge in New York City to sell ya. >:-)

how much?! :cool:

Karel

10-19-2007, 02:38 PM

Dambit! All my monkeys produce is "The Taming of the Shrew." They are no where near coding... :(

Let's try it. Progranism (http://www.progranism.com/) is interesting, so choose a (preferably interpreted, compiling the source would produce far to many errors) language with not so strict rules, create a set of rules (e.g. ad semicolon at the end of line, to become the monkeys God aka Intelligent design :-D) and let it run. Delete those that don't run, let the rest spread. A good start :)

zsouthboy

10-19-2007, 02:49 PM

...i'd like to know if compiz-fusion would be able to run at decent speed with old xpress200m cards.

Heads up - Ubuntu Gutsy works great with this - it automatically sets up an xgl session and everything for you. Use the restricted drivers applet.

200m performance in CompizFusion is excellent.

Uchikoma

10-19-2007, 04:03 PM

Heads up - Ubuntu Gutsy works great with this - it automatically sets up an xgl session and everything for you. Use the restricted drivers applet.

200m performance in CompizFusion is excellent.

So it sets an xgl session up for you automatically...sounds cool...except for one thing.

How are they determining whether to set up an xgl session or use aiglx? If they check for what driver you're using, it'll still try to setup an xgl session even though aiglx is available for our use.

givemesugarr

10-19-2007, 04:46 PM

Heads up - Ubuntu Gutsy works great with this - it automatically sets up an xgl session and everything for you. Use the restricted drivers applet.

200m performance in CompizFusion is excellent.

i don't like xgl, since it's damn slow.... i had from a min of 7fps to 85fps with xgl-beryl running the last time that i tried it. then i had a problem that would have driven me mad at every startup: first it would have tried aixgl, then switch to xgl, then switch to default kwin after about 2 minutes, then switch again to xgl after 1 more minute giving me buggy textures. after that to use xgl i would have to reset kwin and then manually select xgl.
another problem is that i hate gnome (that is more and more awful after the mono-evolution-beagle-gnome integration) so ubuntu for me wouldn't work.
the last problem is that kubuntu is awful. i run gentoo i have a fully customized and optimized system on which i can install everything i want. on kubuntu i couldn't do it. compiling firefox for example it would have been a nightmare (i wasn't able to do it with only qt and no gtk since ubuntu/kubuntu doesn't have a qtdevelop package or at least i wasn't able to find it); i tried also to install ndiswrapper with no luck of making it work.
but this is not the discussion to start a distro war, so i want to close this digression before it starts.
so, reading that the new "answer to the fundamental question driver" is expected to support aiglx i'd like to know if it will really do it for my board too.

sdlvx

10-19-2007, 04:54 PM

I'm scared.

I tried the 8.41.7 drivers, and they worked almost perfectly fine, except for the fact that when the drivers were installed, kicker would crash. Uninstalling them fixed the problem.

There was a big performance gain. I'm just worried that I'm going to end up with a specific bug that's not going to be fixed in this release, and the drivers are going to slip through my fingers again.

The two biggest things I want are that terrible bug that causes massive distortion on the screen until you restart X and AIGLX. The performance gain from 8.41.7 on my mobility x1600 was really good. Everything felt smooth and it all worked great.

I also hope that they fix(ed) the problem with decoration blur on emerald. It's an ATI specific problem, so I'm guessing it's the drivers. The lack of blur really really angers me, specially when I have an 8800 GTS in my desktop that does all of this wonderfully. I think it's just a problem with rendering within the drivers.

mits

10-19-2007, 05:46 PM

Heads up - Ubuntu Gutsy works great with this - it automatically sets up an xgl session and everything for you. Use the restricted drivers applet.

200m performance in CompizFusion is excellent.
xgl on my HD2600 mobility is worse than nonexistant...
performance is like the gpu is completely bypassed, and the cpu is rendering the windows pixel-by-pixel using my granny's algorithms, while undercloked at 100MHz...

heruan

10-19-2007, 06:17 PM

So it sets an xgl session up for you automatically...sounds cool...except for one thing.

How are they determining whether to set up an xgl session or use aiglx? If they check for what driver you're using, it'll still try to setup an xgl session even though aiglx is available for our use.

Gutsy sets up XGL sessions *only* if you manually installed xserver-xgl package. Uninstalling it makes automatic XGL sessions disappear.

daniel of sarnia

10-19-2007, 07:29 PM

i don't like xgl, since it's damn slow.... i had from a min of 7fps to 85fps with xgl-beryl running the last time that i tried it. then i had a problem that would have driven me mad at every startup: first it would have tried aixgl, then switch to xgl, then switch to default kwin after about 2 minutes, then switch again to xgl after 1 more minute giving me buggy textures. after that to use xgl i would have to reset kwin and then manually select xgl.
another problem is that i hate gnome (that is more and more awful after the mono-evolution-beagle-gnome integration) so ubuntu for me wouldn't work.
the last problem is that kubuntu is awful. i run gentoo i have a fully customized and optimized system on which i can install everything i want. on kubuntu i couldn't do it. compiling firefox for example it would have been a nightmare (i wasn't able to do it with only qt and no gtk since ubuntu/kubuntu doesn't have a qtdevelop package or at least i wasn't able to find it); i tried also to install ndiswrapper with no luck of making it work.
but this is not the discussion to start a distro war, so i want to close this digression before it starts.
so, reading that the new "answer to the fundamental question driver" is expected to support aiglx i'd like to know if it will really do it for my board too.

Ubuntu never used beagle, for indexed search it is only now using tracker. The only things installed by default that are mono dependent are f-spot and tomboy notes, which can easily be removed. I also know for a fact that the lead ubuntu developers including Mark Shuttlworth openly hate mono and just include the two apps mentioned because they are exceptional. Ubuntu is pretty much a 100% python and C shop on the gnome side of things.

I also know qtdevelop has been in the main repos at lest since 5.10, when I started using ubuntu. So I don't know where you were looking. Furthermore any problem you have with ndiswrapper is pretty independent of any distro, ubuntu actually will set it up though the restricted driver manager now.

If you just don't like ubuntu that's fine and dandy, but pretty much everything you just complained about is false, that's all.

Morlyn

10-20-2007, 03:20 AM

i don't like xgl, since it's damn slow.... i had from a min of 7fps to 85fps with xgl-beryl running the last time that i tried it. then i had a problem that would have driven me mad at every startup: first it would have tried aixgl, then switch to xgl, then switch to default kwin after about 2 minutes, then switch again to xgl after 1 more minute giving me buggy textures. after that to use xgl i would have to reset kwin and then manually select xgl.
another problem is that i hate gnome (that is more and more awful after the mono-evolution-beagle-gnome integration) so ubuntu for me wouldn't work.
the last problem is that kubuntu is awful. i run gentoo i have a fully customized and optimized system on which i can install everything i want. on kubuntu i couldn't do it. compiling firefox for example it would have been a nightmare (i wasn't able to do it with only qt and no gtk since ubuntu/kubuntu doesn't have a qtdevelop package or at least i wasn't able to find it); i tried also to install ndiswrapper with no luck of making it work.
but this is not the discussion to start a distro war, so i want to close this digression before it starts.
so, reading that the new "answer to the fundamental question driver" is expected to support aiglx i'd like to know if it will really do it for my board too.

Let me first say that I registered to phoronix to answer this post. It's not that big of a deal, I follow the thread and articles with interest here but this kind of bothered me.

Flat out, on an average computer the difference between an 'optimized' gentoo and latest ubuntu is unnoticeable. I've used gentoo for years and I know what I'm talking about, so let's not clutter this thread with false claims and rants.

Karel

10-20-2007, 03:31 AM

So, I guess, no driver for another two days. I don't want to be modded / banned so I won't post any assumptions :)

Anyway, look at the Radeon HD driver, what dedicated developers were able to produce in such a small amount of time - and look at fglrx, which is produced by a huge company (for our money, of course) for several years...

enzobelmont

10-20-2007, 04:26 AM

Let me first say that I registered to phoronix to answer this post. It's not that big of a deal, I follow the thread and articles with interest here but this kind of bothered me.

Flat out, on an average computer the difference between an 'optimized' gentoo and latest ubuntu is unnoticeable. I've used gentoo for years and I know what I'm talking about, so let's not clutter this thread with false claims and rants.

you're not completely right,my friend.

gentoo is years ahead ubuntu in customization, if you're a gentoo user you'll understand it.

sugarr did not said that ubuntu was slower than gentoo, he said that XGL is slow, (we all know that, because xgl is an UGLY hack to xserver).

please be patient.... the hour is coming. :)

sorry my english... ;)

givemesugarr

10-20-2007, 04:30 AM

Ubuntu never used beagle, for indexed search it is only now using tracker. The only things installed by default that are mono dependent are f-spot and tomboy notes, which can easily be removed. I also know for a fact that the lead ubuntu developers including Mark Shuttlworth openly hate mono and just include the two apps mentioned because they are exceptional. Ubuntu is pretty much a 100% python and C shop on the gnome side of things.

I also know qtdevelop has been in the main repos at lest since 5.10, when I started using ubuntu. So I don't know where you were looking. Furthermore any problem you have with ndiswrapper is pretty independent of any distro, ubuntu actually will set it up though the restricted driver manager now.

If you just don't like ubuntu that's fine and dandy, but pretty much everything you just complained about is false, that's all.

that's was my experience some months ago and i had a terrible experience. now i'm having a very pleasant one with gentoo and i think that i'll never switch back to any other versioning distro. i'll just have to compile the new updates when they become stable or to compile them when i need them. and adding stuff is very easily. it was not like that when i used kubuntu. to not speak of configuring issues: i had a terrible headache finding the configuration files. i couldn't find some of them... if i really want a gnome distro that has everything installed and it installs in less than 15 mins or updates in less than 30 i'll chose opensuse. the same thing happened with fedora when i tried to last use it (it was the 5 version if i remember well) which also was awful. to not speak of ubuntu forums or wiki that have a lot of stuff but not the one that you really need, as opposed to the gentoo ones. and in general gentoo solutions work on almost every distro but not ubuntu/kubutu which need different stuff. all this annoying stuff made me hate more and more that distro. and a lot of my friend who used it and who needed to install of to fix something have my same idea. ubuntu/kubuntu is good for windows users who switch to linux and want something that installs quickly and don't use particular things. for example you need to completely update kubuntu to have installed the new knm 0.2 which on my gentoo has been installed form the day it hit the svn, while on kubuntu 6 you would have a pain in the as* compiling it.

Flat out, on an average computer the difference between an 'optimized' gentoo and latest ubuntu is unnoticeable. I've used gentoo for years and I know what I'm talking about, so let's not clutter this thread with false claims and rants.

gentoo 2007.1 amd64 optimized on turion64-mt37 1gb ram/video 128 xpress200m boots in 1 minute 13 seconds into kubuntu and loads all the stuff in autostart.
kubuntu 6 with the same settings and things boots up in one minute 40 seconds. which gives 27 seconds of boot time less for gentoo, which is about 30% faster than the other one.
opensuse 10.2 boots up in 1 minute and 33 seconds, which is also faster than kubuntu (now removed).
windows xp professional 2minutes and 23 seconds (now removed).
if i have eint installed i can boot up in less than 45seconds.
this is a thing that i tried on my notebook. now i'll also try the opensuse 10.3 to see what happens.
maye it's me that i cannot configure kubuntu, but i found out that for my standars is an awful distro and it's based on apt and debian which i personally hate. but as i've said, this is no place to have such discussion.

Anyway, look at the Radeon HD driver, what dedicated developers were able to produce in such a small amount of time - and look at fglrx, which is produced by a huge company (for our money, of course) for several years...
i think that the devs at novell are among the best in the linux world. a lot of good stuff has been produced by them in the last years. remember that at the time linux has started to be installed on desktops there were basically 4 distros: debian, red hat, suse and slackware and suse was among the best 3 if not the best, at least in terms of usability at that time.

piccolotux

10-20-2007, 05:21 AM

gentoo 2007.1 amd64 optimized on turion64-mt37 1gb ram/video 128 xpress200m boots in 1 minute 13 seconds into kubuntu and loads all the stuff in autostart.
kubuntu 6 with the same settings and things boots up in one minute 40 seconds. which gives 27 seconds of boot time less for gentoo, which is about 30% faster than the other one.
opensuse 10.2 boots up in 1 minute and 33 seconds, which is also faster than kubuntu (now removed).
windows xp professional 2minutes and 23 seconds (now removed).
if i have eint installed i can boot up in less than 45seconds.
this is a thing that i tried on my notebook. now i'll also try the opensuse 10.3 to see what happens.
maye it's me that i cannot configure kubuntu, but i found out that for my standars is an awful distro and it's based on apt and debian which i personally hate. but as i've said, this is no place to have such discussion. .
Well...boot times are not so decisive since there are too many variables (services for istance). Don't get me wrong Gentoo is a fantastic distro...I've been using it for years but for the average user with normal needs Gentoo is "too much" :)

kirù

10-20-2007, 05:41 AM

I used gentoo for some months, now I think that people can't waste all the life to compile kde or gnome...
When I used gentoo I spent 99% of the pc time to compile and 1% to use the pc... The speed was not so better than ubuntu, gnome-terminal opened in 1 sec on ubuntu and 1.5 on gentoo! Why do I have to loose all this time?

piccolotux

10-20-2007, 05:48 AM

Let's stop talking about Gentoo (definetely OT) :) I'd prefer to have my new ati driver released on monday :p

HighHo

10-20-2007, 05:52 AM

I agree, only I want them yesterday. :)

Phlogiston

10-20-2007, 06:48 AM

I just wrote a script that autochanges it to lowest setting on startup. I don't game too much yet, so it's suits my needs.

...as for battery life, I get roughly the equivalent battery life as Windows, so I don't see a difference.

You could use rovclock, then you don't need a running Xserver to change the settings.

Trek1701

10-20-2007, 09:29 AM

So, back to the original topic...

Is there a release date for 8.42 driver or the only thing I can do is to cross my fingers and wait?

Trek1701

Michael

10-20-2007, 09:33 AM

So, back to the original topic...

Is there a release date for 8.42 driver or the only thing I can do is to cross my fingers and wait?

Trek1701

Yes, basically. Just keep checking back on Phoronix and you'll know when it's out :)

Trek1701

10-20-2007, 09:36 AM

Thanks Michael!

piccolotux

10-20-2007, 10:24 AM

:D Thanx Michael but we all already know there's a release date (sorry)...the debate is all about WHEN. Am I wrong? :D

container

10-20-2007, 10:26 AM

I guess I will be trying Compiz with AIGLX and all that once this driver comes out. But I'd like to ask you, will I have to turn Compiz off before running a 3d app like a game? I mean, will it be completely usable? I've only used Xgl so far, and I think everyone agrees that it isn't really usable. The reason I ask this is that I've seen the advice of some forums to "turn off compiz" if your apps perform badly or glitch (not ati-specific).

piccolotux

10-20-2007, 10:32 AM

meden

10-20-2007, 11:35 AM

What item(s) are you looking forward most to in this release?

1. Video decoding acceleration (AVIVO support)
2. Video decoding acceleration (AVIVO support)
3. Video decoding acceleration (AVIVO support)
4. Reliable power management
5. AIGLX (eyecandy is NOT SO important, men...)

yoshi314

10-20-2007, 11:36 AM

how about stability for a change?

Danyul

10-20-2007, 11:56 AM

If this driver comes out with something ridiculous like the water mark of a couple of earlier releases I think I'm just going to laugh and cry at the same time.

But the major things I would like to see come in to these drivers are:

1. Better dual monitor support. I have a lappy with an x1400 (they brought out the nvidia option after I already bought it.) however I can change the card. But I also hook a 27" monitor up to this thing and I want to use its native res of 1920 x 1200 while still being able to drag programs from one screen to another (and have iceweasel open on both screens). My lappy does a 1680x1050 res on its screen.

2. 2.6.23 support. I have a nice hot kernel built but if I want to use it I have to just enjoy its power on the console.

I've been watching these forums for months but never commented but if these 2 things were included i would be satisfied.

Of course on the ATI time scale I would probably have to ask that they include support for kernel 3.0 support now to make sure they get it on time.

Karel

10-20-2007, 04:16 PM

Also, wann kommt die Wunderwaffe?

Sorry, just couldn't find a better for the driver name - it should be 8.42.4 Die Wunderwaffe. Go read some WW2 history to figure why :)

thoemy

10-20-2007, 06:35 PM

Wunderwaffen is German for "wonder weapons", and was a term assigned during World War II by the Nazi propaganda ministry to a few revolutionary "superweapons". Most of these weapons reached the combat theatre too late, and in too insignificant numbers (if at all) to have a military effect. Even if they did, the "Wunderwaffen" mostly failed to fulfill their propaganda-inflated capabilities. Most soldiers, and many civilians, realized that.
I think many people will not like that...

But I'm all for naming the releases after some random German word :)

Malikith

10-20-2007, 07:57 PM

I think many people will not like that...

But I'm all for naming the releases after some random German word :)

How about if we just call it 8.42 Nuke Edition, it could mean either good or bad, it could mean either it'll blow you away in amazement, or it'll blow your computer off the face of the Earth or cause you to want to nuke the card from orbit. Hopefully we'll know very soon.

link2zelda

10-20-2007, 09:55 PM

How about::
8.42.x Beowulf Edition: Uber-Xtreme, the hero of the Geats, battling Grendel, the beast of no AIGLX, and defeating Grendel's mother, the need for XGL to compiz/beryl desktop effects.

Kano

10-20-2007, 11:20 PM

For any driver: Always half broken Edition.

daniel of sarnia

10-21-2007, 01:18 AM

just once I'd like an "everything works and runs about as fast as on windows Edition."

OOOhhh I don't even have and idea of how fast my X1900XTX is because I can't stand working on windows, and the linux drivers are lame. Wired thing though, when I tried 8.41.7 I stared up quake 4 and maxed out the settings. I then started a game and there was sum corruption although stuff seemed to look better. In the background the black smoke rising into the air on the first level was make up of about a thousand little grey cubes. So I ended up rolling back the driver because it obviously was not stable with my card.

Also, does any one know if with the new driver that with my X1900XTX I'd be able to play UT3 @ 1680x1050 with high settings and get decent FPS? I'm looking forward to that game so much. Michael maybe you have an idea of how well it'll run? ;)

Malikith

10-21-2007, 01:22 AM

daniel of sarnia

10-21-2007, 01:39 AM

I hope so.

Hey also when I run quake 4 or the ET demo, even with the 8.40.4 driver I get a lot of corruption. I wish I know how to take in game screen shoots to how you what I mean. Does anyone else have problems like this I'm running the most up to date client too.

Maybe the new driver will solve this. Come on all singing all dancing second coming of jesus driver, keep your toes crossed.

MagicTempest

10-21-2007, 01:59 AM

The same here. When I choose a resolution higher than 630x480 only an area about the size of 640x480 is rendered correctly but the rest is gibberish.

d2kx

10-21-2007, 04:37 AM

daniel of sarnia,

UT3 really does not need insane graphic cards. It should work with High, maybe not the highest but well let's see. That's what we see with the Windows demo.

Ontopic:

8.42 - The AIGLX-is-here driver! (currently broken for X.org Server 1.0-1.4 edition)

Kano

10-21-2007, 04:43 AM

The UT3 Win demo has no higher textures than medium, setting to high does not change anything. You bench wrong.

heruan

10-21-2007, 04:44 AM

Will tomorrow be The Day?

daniel of sarnia

10-21-2007, 05:06 AM

if does not I'll cry... again and ever day until it comes out. If it sucks I'm taking a hammer to my video card and getting and NV.

enzobelmont

10-21-2007, 05:21 AM

i'm in a crisis point:

if i could change my videocard i would did it a long time ago, but unfortunately it is in a laptop.

if THIS release is not barely bearable, i will put my laptop on ebay.

i'm really tired of a company that can not give their customers decent drivers, in few hours(maybe days) ATI could lose one more client.
it is up to THIS DRIVER RELEASE.

please ATI do no let your customers down/go.

sorry my english.

Phlogiston

10-21-2007, 05:45 AM

Is it true that the driver will not support 2.6.23 kernel?

Kano

10-21-2007, 05:51 AM

For 2.6.23/32 bit there should be no problem to support drivers from 8.40.4 till latest with 2 patches. If needed i could even write a script to update the installer for non debian/ubuntu users. Basically this it not needed, but less complicated to execute the installer when patching is required. There you could use my fglrx script directy which supports 2.6.23 already.

Phlogiston

10-21-2007, 05:54 AM

Kano

10-21-2007, 06:00 AM

For the drivers >= 8.40.4, from the sabayon ebuild:

http://svn.sabayonlinux.org/overlay/x11-drivers/ati-drivers/files/ati-drivers-8.40.4-warnings.patch
http://svn.sabayonlinux.org/overlay/x11-drivers/ati-drivers/files/ati-drivers-2.6.23.patch

d2kx

10-21-2007, 06:06 AM

Late Kernel and X.org support is -the- biggest issue that I always am experiencing with fglrx.

ivanovic

10-21-2007, 07:14 AM

Hmm the patches are for the kernel or the driver? I'm using gentoo, so I guess they will include the patches too.

If you had a look at sys-kernel/vanilla-sources/files/ you would see that those patches *are* included in gentoo. In short: fglrx should work with a 2.6.23 kernel on gentoo (at least with a vanilla kernel).
I would assume that the same patches will be applied to the .42 fglrx release, too. At least they do apply cleanly to 8.41.7 release. I am not using kernel 2.6.23 yet, but the patches are applied when emerging so I assume that it does work nicely.

No idea if those patches will work with gentoo-sources, too. But I'd guess they will.

d2kx

10-21-2007, 08:05 AM

I am running sidux with Kernel 2.6.23.1 and Xorg 7.3 using Kanos fglrx script and it runs perfectly. I've heard about 64Bit issues with the Kernel patches, but haven't tried myself yet.

givemesugarr

10-21-2007, 09:06 AM

Hmm the patches are for the kernel or the driver? I'm using gentoo, so I guess they will include the patches too.
wait for the driver to hit portage (i think at the end of the current day of release or at most the end of the day after the release) or sabayon. remember just one thing: try not syncing more than once a day. i'll suggest that you take a look at the forums and see it it hits the repo. or try watching the websvn of sabayon and see if it hits sabayon first. to use sabayon as a source for ebuilds look around the forums on how to add it as an overlay. of my experience latest ati-drivers have hit sabayon first and worked without flaws on my board, which is a nasty one.
i don't know of problems with 2.6.23 and 64bits, but i'll let you know when i try it out. i had problems with 64bits and 8.41 not compiling and with 8.40 compiling with sandbox problems (there is a way to correct it if you look in the forums) but i didn't yet tried 8.42 since it's not out. i'll let you know better if there are problems when i try it out.

Danyul

10-21-2007, 11:09 AM

if i could change my videocard i would did it a long time ago, but unfortunately it is in a laptop.

Enz, what type of laptop do you have? Because I know with mine (a modified dell 6400) I am able to remove the graphics card and replace it.

I do hate pulling the laptop apart though due to the number of screws holding the thing together.

enzobelmont

10-21-2007, 11:53 AM

Enz, what type of laptop do you have? Because I know with mine (a modified dell 6400) I am able to remove the graphics card and replace it.

I do hate pulling the laptop apart though due to the number of screws holding the thing together.

its a compaq presario v2000, x200m, ah. my sad life... i should spent some more buckets to get something better, if i should know.

sorry my english.

givemesugarr

10-21-2007, 01:30 PM

x200m
this is an igp board which means that there's no way of removing it for what i know.

enzobelmont

10-21-2007, 04:19 PM

this is an igp board which means that there's no way of removing it for what i know.

maybe i will be using a hammer to remove A T I from my life forever.

even if i lose my $900 laptop (that was the cost of it a year ago here in Mexico)

on THIS release depends my decision.

givemesugarr

10-21-2007, 04:45 PM

maybe i will be using a hammer to remove A T I from my life forever.

even if i lose my $900 laptop (that was the cost of it a year ago here in Mexico)

on THIS release depends my decision.

well, with 8.40.4 works quite well if you have atieventsd loaded as a service at boot, the opensource driver has aiglx support and about 1,2 times in fps and supports aiglx (but without compiz which still blurs a little) with the last xorg-server and xf-86-driver-ati. i've not been experiencing issues for 3 months now. and 8.42 promises to give a great boost and aiglx support for this board very similar to the windows one.

enzobelmont

10-21-2007, 04:52 PM

well, with 8.40.4 works quite well if you have atieventsd loaded as a service at boot, the opensource driver has aiglx support and about 1,2 times in fps and supports aiglx (but without compiz which still blurs a little) with the last xorg-server and xf-86-driver-ati. i've not been experiencing issues for 3 months now. and 8.42 promises to give a great boost and aiglx support for this board very similar to the windows one.

i know the facts about performance in x200m;

in few hours we'll see if wait was worth.

givemesugarr

10-21-2007, 05:06 PM

in few hours we'll see if wait was worth.

ahahahahaah.... do you really expect them to release it early in the morning?!?! i expect a release for after tomorrow in the late morning and the driver hit in portage Wednesday morning.
let's see if i'm right...

Uchikoma

10-21-2007, 05:27 PM

enzobelmont: go ahead and smash it with a hammer as I don't believe this release will miraculously solve all problems...just some of them (and add AIGLX support)

I doubt they'd release in the early morning too.

Michael

10-21-2007, 05:29 PM

It won't be out tomorrow morning...

koolmanoncampus

10-21-2007, 05:52 PM

tomorrow afternoon? haha

toxigenicpoem

10-21-2007, 07:42 PM

i know the facts about performance in x200m;

in few hours we'll see if wait was worth.

Performance boost in 8.41 was terrific for the x200M, i can't wait for the stability, and extra goodies of 8.42

Uchikoma

10-21-2007, 08:31 PM

Performance boost in 8.41 was terrific for the x200M, i can't wait for the stability, and extra goodies of 8.42

That's good to hear. I wonder what it'll do for my x1400 and x1800 respectively.

edged

10-21-2007, 08:39 PM

I wonder if ATI will "fix" the Catalyst Control Center to be compatible with X.Org distros that use XCB:
amdcccle: xcb_xlib.c:52: xcb_xlib_unlock: Assertion `c->xlib.lock' failed.
Aborted

There is a work-around for this (an environment variable), that supposedly works for other apps that have locking problems:
export LIBXCB_ALLOW_SLOPPY_LOCK=1

It doesn't work for me, although I wonder if it works for anyone else.

enzobelmont

10-21-2007, 10:02 PM

Uchikoma

10-22-2007, 01:23 AM

what I expect from 8.42 is:

- a slighly better performance than 8.41 (some more FPS).
-AIGLX support.
-stability.

is that impossible??

optionally (i think this will not be included in this release):
-latest xorg support.
-latest kerrnel support.
-multiple screens support.
-a working control center.
when i said few hours, i was talking of 40 hours approx (maybe less).

sorry my english... ;)

It might be impossible :) At least for this upcoming release anyways. If they're really getting their stuff in gear, we'll see results, but we won't see 100% uberness and stability right away.

You know what they say, 'better late than never'

dopehouse

10-22-2007, 02:07 AM

I think the driver will be released between 8pm and 11pm UTC+2 ;)

And I hope that the performance will be improved improved improved. The ETQW demo is not playable where the Windows demo is running with high settings on my x850xt.

And I also hope that GoogleEarth is running too, again.

givemesugarr

10-22-2007, 02:45 AM

Performance boost in 8.41 was terrific for the x200M, i can't wait for the stability, and extra goodies of 8.42
yeah, just on 32bit distros. on 64bit you cannot even compile it unless you have gcc 3 installed (which is pretty impossibile).

-latest xorg support.
-latest kerrnel support.

ahahah... dream on little boy... the latest kernel was released after the feature add of this release was finished. the only thing that you could hope is xorg7.3 and xorg-server 1.4 support which were released about a week before of the latest kernel and if i'm correct that was still during the feature add phase.
but for this assertion to be right we need someone to confirm it. unfortunately all devs have ndas with amd/ati and it's quite probable that we'll have to wait for the driver to be released to have such info from them.
anyway, i just have to say that in the last period the driver had quite some progress. and with the oss radeonhd project things seem to go for a better support.

Kano

10-22-2007, 02:48 AM

givemesugarr

10-22-2007, 02:58 AM

You can not compile on 64 bit? Thats wrong, tested it always with current 64 bit live cds and it always worked. Maybe your kernel was compiled with gcc-3.3, then you need it for the kernel module too. Check:

cat /proc/version

The same compiler should be used for additional kernel modules. If your kernel was compiled with an old compiler then recompile it or install old compiler!

nope. i cannot install 8.41.7 on amd64 arch with gentoo and gcc4.1.2 installed. the kernel is new and compiled with 4.1.2 and i don't even have gcc3 installed if not for the libstdc++-3 which is a runtime dependency for the driver. i always get a libstdc++ error. it's a very annoying issue!!!! now i'm currently going with oss drivers aiglx but no compiz/beryl since it doesn't work well with x200m and the textures are pretty bad sometimes.