Shader problems with ATI Radeon drivers [Solved]

[Zieman]

yeah, sorry about the gl_TexCoord[0].st thing, that was a really daft bug to have introduced into the shaders,
hopefully i'll have the normal mapped shipset updated soon with fixed shaders, the new anaconda shaders might interest you Frame regarding your registration oxp, the decal code has been moved into a function in the shader, you just need to call this function multiple times (with updated decal position & rotation info each time) to add multiple decals to the ship, this cuts down on a lot of duplicated decal related code in the shader when adding multiple decals

Thanks , mistakes happens, all the time.. was it not for your work, a lot of the shaderfied oxps might never have happened...

That is going to be the next update to the Registration ID OXP.. as rotation seems a requirement to place the ID correctly...

Alas, looks like many OXPs now share the shader bug - us ATI users have to wait for fixed shaders (if using catalyst 9.11 - 10.3) for at least:
Anarchies
Freaky Thargoids
Lave Academy
NeoCaduceus
Neolite ships
PAGroove Aurora
PAGroove Stations
Ramon Anaconda
Shady Blackmonks
Feudal States
Vector
Your Ad Here

[Killer Wolf]

Bah.

mate gave me the demo disc of Crysis to give my new quadcore something to get its teeth into. it works great and looks stunning, but there was a couple times where the game would just go loopy, rapid changing of gun settings, the player's POV zipping about etc. didn't think much at the time, was half thinking the pre-release thing might not been as stable. Did it again last night tho, and in between fiddling, i've sussed it : it's that stupid fkn prob where the vid stuff crashes, flittering the mouse all over the screen etc etc. most of everything i do is stable, so i'm wondering if there's some graphical thing in the game that just causes the card to go haywire.
tell you what tho, this is the last time i go AMD/ATI : didn't have a pic of bother w/ intel/nvidia, but i seem to have had nowt but annoyances w/ this setup. if they can't even get their own utils to sync w/ their own brand cards and processors, there's not much hope for them.

[Cody]

<sigh> It's new driver time again at ATI/AMD... Catalyst version 10.4 is here.
Somehow, I don't have that much enthusiasm for installing it, but I'll try it out tomorrow (maybe).

[Cody]

Catalyst version 10.4, driver version 8.723 now installed... and so far, so good.
We'll see how it behaves over the next few days, which will involve some heavy game play.
More deep immersion... the things one has to suffer!

[Star Gazer]

Ahhh... what a hero! The suffering artiste... ...it's a dirty job, but someone has got to do it...

[Palmski]

Catalyst version 10.4, driver version 8.723 now installed... and so far, so good.
We'll see how it behaves over the next few days, which will involve some heavy game play.
More deep immersion... the things one has to suffer!

Does it fix the problems with shaders do you know? With CC 10.3 the Con Stores for example look flat and drab which I understand is an ATI driver problem.

[Cody]

Hi Palmski

As I understand it, it’s the shader code that causes the problem. Later ATI drivers (9.11 onwards) deal with it differently, hence the problem. I don’t think that 10.4 will help any, but it may be worth trying. The more feedback the better. Griff tweaked the code for his ship set, which is all I have knowledge of, I’m afraid.

If you feel like experimenting, try rolling back to Catalyst 9.10 (loss of performance is minimal), and see if the Constores etc look any better.

[Cody]

Something that I’ve noticed already… Catalyst 10.4 (like 10.2 and 10.3) stutters and scrambles the display briefly when starting Oolite (even displaying the config screen very briefly before the spinning Cobra appears on some occasions). It’s just long enough to give you a moment of slight worry, then it’s fine. The odd thing is, using the ‘no-splash’ thingy seems to cure the problem, and I’ve never had to use the ‘no-splash’ before. No black screen/VPU recover events so far, though.

Otherwise, it appears to be fine, as far as the starship ‘Rolling Thunder’ is concerned. That’s the trouble with this beautiful game… I launch, intending a series of short flights, exits and restarts, just to check out the new GPU driver, get jumped by some bandits, and I’m away… ‘immersion’ takes over, I’m off across several systems, my kill count has risen, and a couple of hours have flown by.
Oh well, I suppose I’d better get back out there.

[Diziet Sma]

That’s the trouble with this beautiful game… I launch, intending a series of short flights, exits and restarts, just to check out the new GPU driver, get jumped by some bandits, and I’m away… ‘immersion’ takes over, I’m off across several systems, my kill count has risen, and a couple of hours have flown by.
Oh well, I suppose I’d better get back out there.

I'd like to thank you, El Viejo, for all the effort and suffering you put yourself through, as the forum's resident ATI driver tester/guru..

[Palmski]

I had a couple of hours of intensive testing last night too and could see no differences between 10.3 and 10.4. I haven't experienced the stuttering/scrambling though for any of the Catalyst versions.

Did I see somewhere that the shader code has been amended for 1.74 to take account of the ATI drivers? If so is there any rough guesstimate as to when a release might take place? I promise you I won't hold you to it. Just wanted to get a feel if it was worth messing about with rolling back drivers - something I have found to be a somewhat flaky process when I have done it in the past.

[Griff]

Did I see somewhere that the shader code has been amended for 1.74 to take account of the ATI drivers? If so is there any rough guesstimate as to when a release might take place? I promise you I won't hold you to it. Just wanted to get a feel if it was worth messing about with rolling back drivers - something I have found to be a somewhat flaky process when I have done it in the past.

unfortunately the faulty shaders are all inside oxps, the built in Oolite shaders are OK (as far as i know) so all the built in ships and any oxp ships using materials will display properly, i'm sure i saw a list of oxp's that needed fixing posted here on the forum edit: heh, they're listed at the top of the page

[Palmski]

I should have seen that myself.

I can confirm that the latest PAGroove Stations since updating look fine now. I have had no problems with the Neolite ships though, they all look as they should AFAIK.

Of the rest I've only got Your Ad Here and Feudal States, the former definitely still is broken and I didn't make it to a Feudal system last night so couldn't say. Out of interest should the Random Hits bars look "Griffy"? They are looking largely like grey silhouettes for me so might need adding to the list.

How straightforward is it to rectify ie could I do it as a Java programmer with no graphics experience armed with Notepad++? I would happily have a stab at getting it working locally if I had a vague idea of what was involved.

[Griff]

Although i haven't tested this, it might be as simple as changing any references to gl_TexCoord[0].st to gl_MultiTexCoord0.st in all the fragment shaders in these oxp's.
eg, change the line

Code: Select all

 
vec2 texCoord = gl_TexCoord[0].st;

to

Code: Select all

vec2 texCoord = gl_MultiTexCoord0.st;

this might work for the moment, although all these old shaders really need to be fixed up and brought to to date (they use a macro to work out the lighting calculations needed and they should be using a 'function' to do this instead, i'll have a look to see if i can sort something out by comparing these old shaders with the newer ones and transfering bits across.

[Palmski]

Cool thanks, I'll have a stab at that this weekend. I'll let you know...

[ADCK]

Although i haven't tested this, it might be as simple as changing any references to gl_TexCoord[0].st to gl_MultiTexCoord0.st in all the fragment shaders in these oxp's.
eg, change the line

Code: Select all

 
vec2 texCoord = gl_TexCoord[0].st;

to

Code: Select all

vec2 texCoord = gl_MultiTexCoord0.st;

this might work for the moment, although all these old shaders really need to be fixed up and brought to to date (they use a macro to work out the lighting calculations needed and they should be using a 'function' to do this instead, i'll have a look to see if i can sort something out by comparing these old shaders with the newer ones and transfering bits across.

Slightly off-topic,
I use neolites shaders, and they've been mentioned as those being effected.
I'm tired of having to deal with complicated shaders since i am a lazy person who likes things simple. So...
I know illumination and shinyness can be done with materials, but what about normal maps?
I only ask because if normal maps can be done with materials, I wouldn't even need shaders anymore (except for animated stuff)