XiG proposition?

Printable View

12-25-2007, 04:57 AM
RichardB

XiG proposition?

Hi,

I'm a long time Linux user, with an ATI X1400 chipset in my laptop, and have just rolled my system back to 8.35.5 (last FireGL release, should be well tested...) after the mess with 7.12. However, I've also used (bought) XiG's Acclerated X with success on an old 9250 series card, and I'm wondering what the feeling is about pay-for graphics drivers (or X servers) that actually work.

If XiG had the R500 and newer docs that the RadeonHD project does, would anyone (apart from me...) be prepared to pay for their product?

Cheers,
Rich

P.S. No affiliation with XiG, I've just used their stuff before, and am sick of the drama with ATI drivers.
12-25-2007, 07:59 AM
Svartalf

Quote:

Originally Posted by RichardB

If XiG had the R500 and newer docs that the RadeonHD project does, would anyone (apart from me...) be prepared to pay for their product?

Based on a current reading of what the Desktop lineup currently supports, they don't have anything in hand that'd be better than what's already supported (and decently enough at that) under the free software DRI drivers. Now, if they DID get technical specs, they might be able to come up with better than adequate drivers, but unless they pour the coals to things, they won't get it done much faster than the FOSS teams will if we get what AMD's promising us in the next 6 or so months. Seriously.

IF they had something and IF it worked as well as their site claims their stuff does (which is closer to the truth with them than many companies...) it MIGHT be worth the $89 for a desktop or laptop edition driver purchase. But only as long as AMD keeps producing unstable drivers for us and we don't have some FOSS driver support that is close to the same.
12-25-2007, 08:38 AM
RichardB

Quote:

Originally Posted by Svartalf

Based on a current reading of what the Desktop lineup currently supports, they don't have anything in hand that'd be better than what's already supported (and decently enough at that) under the free software DRI drivers. Now, if they DID get technical specs, they might be able to come up with better than adequate drivers, but unless they pour the coals to things, they won't get it done much faster than the FOSS teams will if we get what AMD's promising us in the next 6 or so months. Seriously.

IF they had something and IF it worked as well as their site claims their stuff does (which is closer to the truth with them than many companies...) it MIGHT be worth the $89 for a desktop or laptop edition driver purchase. But only as long as AMD keeps producing unstable drivers for us and we don't have some FOSS driver support that is close to the same.

Right, it's not something they current have on offer, however from what I understand about docs supplied for the RadeonHD driver, AMD is more forthcoming with documentation these days. That said, if supplied with docs I wouldn't like to wager on their development time-lines to support a new chipset. Maybe they've got very modular code, maybe not...

Look, don't get me wrong, I'm a big fan of FOSS development. I'm also fairly pragmatic; ATI drivers have been a bit flaky as long as I've been using them (years), and the FOSS drivers took a while and never quite seemed to be 100% feature complete... certainly RadeonHD looks like it has a LONG way to go (wiki says no 2D accel, no 3D accel, no XV, etc, etc). And once it's working, how long before it's working right?
12-25-2007, 09:12 AM
Svartalf

Quote:

Originally Posted by RichardB

Look, don't get me wrong, I'm a big fan of FOSS development. I'm also fairly pragmatic; ATI drivers have been a bit flaky as long as I've been using them (years), and the FOSS drivers took a while and never quite seemed to be 100% feature complete... certainly RadeonHD looks like it has a LONG way to go (wiki says no 2D accel, no 3D accel, no XV, etc, etc). And once it's working, how long before it's working right?

Actually, the FOSS drivers, if you're referring to the R100/R200 or Intel chipsets is fairly complete. If you're referring to the R300/R400 drivers, heh... To say that they're not complete and to use the timeline that we've seen for them is not to be fair to the hard work people put into them. They did what they've done WITHOUT ANY INFO WHATSOEVER. If you're talking about a few bits and bobs being missing in the R100/R200 driver- that's because we didn't get the info and as you can see with the R300/R400 reverse engineering isn't a simple or easy task.

If we've got decent information, which it appears that AMD is trying to get us, it won't take but about 6 months for us to get something better than adequate on those parts.

The truth of the matter is, it'd take XiG the same or more amount of time to get something from when the tech data is released- for something that you'd have to pay lots of money for on EACH machine you use it on. Being pragmatic is nice, but when I tell you what I tell you, I'm being as pragmatic as you are. You just don't have all the info I do. :D
12-25-2007, 07:04 PM
RichardB

Quote:

Originally Posted by Svartalf

Actually, the FOSS drivers, if you're referring to the R100/R200 or Intel chipsets is fairly complete. If you're referring to the R300/R400 drivers, heh... To say that they're not complete and to use the timeline that we've seen for them is not to be fair to the hard work people put into them. They did what they've done WITHOUT ANY INFO WHATSOEVER. If you're talking about a few bits and bobs being missing in the R100/R200 driver- that's because we didn't get the info and as you can see with the R300/R400 reverse engineering isn't a simple or easy task.

If we've got decent information, which it appears that AMD is trying to get us, it won't take but about 6 months for us to get something better than adequate on those parts.

The truth of the matter is, it'd take XiG the same or more amount of time to get something from when the tech data is released- for something that you'd have to pay lots of money for on EACH machine you use it on. Being pragmatic is nice, but when I tell you what I tell you, I'm being as pragmatic as you are. You just don't have all the info I do. :D

I'm a software engineer who has written driver level code and debugged plenty of software... so I appreciate the difficultly of reverse engineering this stuff; I'm not belittling your efforts. And yes, equally skilled developers working within equally well designed frameworks would take equally long to bring something stable to end-users.

Lately however, I've been working on 3D VR-type apps (3D graphics and 3D audio), and whether it's for XNA on Windows or for OpenGL on Linux or whatever, the quality of the graphics drivers matter. I don't mind paying for software (yes, for each and every ONE of my ATI-based machines) if it works correctly (and quickly, as a bonus), which the free software Xorg drivers (both ATI and OSS) do not always do. Problems with TV-out on an old 9250 card prompted my previous purchase, and current rendering problems on my laptop have prompted this thread.

When posting this thread, I was looking for input from from other end users - not from competing developers. Good luck with your work - Like most folks, I'll use which ever works best, whether it's FOSS, closed-source, or payway.
12-25-2007, 10:25 PM
Svartalf

Quote:

Originally Posted by RichardB

When posting this thread, I was looking for input from from other end users - not from competing developers. Good luck with your work - Like most folks, I'll use which ever works best, whether it's FOSS, closed-source, or payway.

Heh... I am not currently developing drivers. I'm developing visualization code, game code, etc. and I'm one of those end-users you refer to more than anything else. I've just got a solid handle on what all it takes to make the code come together because of the past work I've done in the driver space, especially 3D drivers.

However, having said this, you've got to be careful about making comments like the one you made. The very people you're commenting about, "competing developers", are the very users you're talking to in many cases, myself included. Making comments like that step on the toes of the people that make your entire platform possible in the first place. But, that's a different discussion, so I'll let it go at just this in-passing comment about it...

In the end, I don't care who makes the drivers- right now I'm using NVidia in most of my stuff because of all the fun AMD is having with us on their products. I prefer to have FOSS drivers for things, but I'm a pragmatist and use what just simply works right now. If FOSS drivers do that (R200, for example...) then that's what gets used. If it's closed source drivers, then that's what gets used.

If XiG was realistically in a position to produce the drivers, I might be interested even though they're very expensive. They're not in such a position as best as I can tell (Otherwise, they would happen to have support for far more than the R200 series chipsets...). Now, I could be mistaken about that, and I would be pleasantly surprised. However, if they don't have the R300/R400, R500, or R600 programming info at this time, they're not going to get there ANY faster than we're about to get there with the FOSS drivers. Honest.

You can view that as being comments from a "competing developer" but you'd be DEAD WRONG on that one. You're just barking up the wrong tree if that's what you think. Now, having said this, you've claimed no affiliation with XiG- do YOU know something about their being able to offer working drivers in the near future that you're not telling us?
12-26-2007, 03:22 AM
RichardB

Quote:

Originally Posted by Svartalf

Heh... I am not currently developing drivers. I'm developing visualization code, game code, etc. and I'm one of those end-users you refer to more than anything else. I've just got a solid handle on what all it takes to make the code come together because of the past work I've done in the driver space, especially 3D drivers.

However, having said this, you've got to be careful about making comments like the one you made. The very people you're commenting about, "competing developers", are the very users you're talking to in many cases, myself included. Making comments like that step on the toes of the people that make your entire platform possible in the first place. But, that's a different discussion, so I'll let it go at just this in-passing comment about it...

In the end, I don't care who makes the drivers- right now I'm using NVidia in most of my stuff because of all the fun AMD is having with us on their products. I prefer to have FOSS drivers for things, but I'm a pragmatist and use what just simply works right now. If FOSS drivers do that (R200, for example...) then that's what gets used. If it's closed source drivers, then that's what gets used.

If XiG was realistically in a position to produce the drivers, I might be interested even though they're very expensive. They're not in such a position as best as I can tell (Otherwise, they would happen to have support for far more than the R200 series chipsets...). Now, I could be mistaken about that, and I would be pleasantly surprised. However, if they don't have the R300/R400, R500, or R600 programming info at this time, they're not going to get there ANY faster than we're about to get there with the FOSS drivers. Honest.

You can view that as being comments from a "competing developer" but you'd be DEAD WRONG on that one. You're just barking up the wrong tree if that's what you think. Now, having said this, you've claimed no affiliation with XiG- do YOU know something about their being able to offer working drivers in the near future that you're not telling us?

I'm not interested in an ITG pissing match with you, Svartalf, so I'll refrain from quoting your use of "we" and "us" in reference to the discussion on current and previous development.

If you're not involved with the current R5xx efforts then I'll apologise for assuming you were. However, for someone without an interest in the current Xorg development, a current Nvidia user no less, you've sure got yourself wound up over me asking if other dissatisfied users would look at an alternative.