[Gregster2k]

CAUTION: Please don't turn this thread into another "Roll Call, who has this problem" type thread, we already had one of those for this problem, search for "breaks Shinemapping!" original bug report thread posted by myself, it has quite a few ppl who also confirmed this bug's existence, even on NON-9800 Pro cards! There was talk of informing ATI at the time. But if ATI WERE informed, then ATI are clearly sitting on their butts regarding this problem.

Video Card & Primary Issue:
- ATI Radeon 9800 Pro 128MB: FSO/SCP Shinemapping implementation completely fails on this card with any driver version above Catalyst 4.4. This does not affect glow mapping. A previous thread was already made about this issue (see the above note).

Secondary Issue:
- The Catalyst 4.4's, while allowing the SCP to have working shinemapping, are killing performance in all my other games. And while newer drivers speed up my other games they destroy the SCP's shinemapping implementation.

Cause:
- Any ATI Catalyst driver version higher than 4.4 and certain ATI graphics cards, including ATI Radeon 9800 Pro.

Solution...no..wait...not a solution...a workaround...#1
Repeatedly swap out my drivers with a driver swap utility of some kind or even manually depending on whether i want SCP or standard games. Of course, this is a ridiculous way to deal with the problem. I still will get performance issues in other games when using the 4.4's, higher-than-4.4's will mean broken SCP shinemapping implementation, and all this driver swapping involves manual reset of all video card setting presets and monitor refresh rate timings for every resolution it can support, also will reset some device resources in the process of the repeated reinstallations which could cause the system to become unstable(!!!).

Solution 2
-==***Requesting a Permanent Solution Status Update***==-
Well, clearly ATI isn't going to accidentally fix this issue and I doubt they are going to if we already asked them to or not. They don't have any reason to care about the SCP, after all its not owned by a corporate entity possessing tons of money, it stands alone and free as an open source project. This bug does not occur on all ATI video cards but has been confirmed to occur on more cards than just the 9800 Pro (see original thread for that information). All other games I play have perfectly functional shinemaps with the latest Catalysts...

Even though the new drivers have brought this problem to the surface, we can't know whether or not it is ATI's fault alone -- it could be the SCP only, after all, FSO is the only shinemapped game the new Catalysts break shinemapping on. And if it is a bug with FSO, then even if ATI fixes their drivers for us, as unlikely as that is, the bug will be lurking, waiting to spring on us in the form of bad framerates or future problems. In addition, anyone first starting out with the SCP will be discouraged if they find they have to DOWNGRADE their drivers to fix the problem. Or, they may think shinemapping doesn't really exist...just kidding, LOL

Whoever tries to take on the (admittedly daunting) task of fixing this horrendous bug has to consider every possibility and act upon it:
- It could be the SCP alone...(I believe this to be the most likely situation here. there is evidence to support this, for one, catalysts that work on one card type DON'T work on another!!!)
- It could be ATI alone...
- And, it could be BOTH at fault here (ATI drivers simply causing an SCP *bug* to express itself, for example. i think this is the 2nd most likely scenario).
The only solutions available, if ATI can't help us on their end (and its better to assume the worst, so let's assume they will NOT help us out)...
...are either the squashing of a potentially-existent hidden bug in the current shinemap implementation, OR, the addition of alternative "ATI fixed" SCP shinemapping code to FSO (best solution IMHO if no others available)

All in all, if any progress has been made already toward fixing this bug, how close are we to seeing a usable bugfix? And if no one is currently trying to fix this, how soon can someone get on it?

Do it for Lightspeed...I really appreciate that guy's efforts for shinemapping FreeSpace ships, and his hard work will be WASTED on people who have my situation. Sure, I'll be able to *install* the shinemaps, *seeing* them is another matter...

Please??  

Sigh...all this said, I'm now going to install the latest Catalysts, enjoy the multiplayer games I play, and wait for this bug to be squashed. I can't stand the sight of FSO without shinemapping, I joined here when it already had it and I can't imagine FSO without it now. This driver upgrade is a necessary evil to FSO for the greater good of my system. If the shinemapping isnt fixed, oh well, goodbye FSO for me, shinemaps are the sole reason I use FSO besides the speed increase, polycount capabilities, multiplayer...

*loads up ati.com's downloads page in silence*
------------------

UPDATE: I just reread this forum and discovered someone with 4.4's has(had?) the problem...that's odd.

http://www.hard-light.net/forums/index.php/topic,25987.0.html

They, however, are using a different ATI card. Whatever this problem is, different drivers cause it on different cards, apparently. And some ATI cards may not work with shinemapping at all. I am willing to bet though that they all have one unified cause, whether it be a problem in ATI's drivers, or, the actual SCP FSO project code itself. Seeing as the person above had 4.4 and was experiencing teh problem, it makes me lean more toward believing that  the SCP code is what is doing all of this, and that it needs to be reworked especially and alternatively for the ATI users here.

[Bobboau]

I'm sure were not the only people affected by this, and if we were ATI isn't going to do a damned thing for a small open source project that has, maybe, 100 users.
this problem will be fixed when we move on to the programable pipeline rather than relying on the fixed function as we have been for some time.

[StratComm]

I'll chime in here to say that it's an issue with the new cards as well, so it's not something that I expect will be getting any better any time soon.  And oh how I'd love to see shines on my X800XT, since I can't get FSO to lose performance except when it hits one of its crashes.

[Woolie Wool]

*considers whether shinemaps or Doom 3 performance is more important*

Shinemaps or Doom 3? Shinemaps or Doom 3? Ummm uhhh uhhhh....

Oh **** it. Decent framerates in games like Doom 3 is worth a temporary loss of specmapping.

[Overlord]

Too bad you guys have no Geforce-Card, uh?

[Turambar]

too bad bobbau wont write any nvidia optimization thingys into any code

[Overlord]

No need, cause it runs pretty fine on my comp.

[xenthorious]

Guys, before you have a cow and give birth to a baby that will end up killing you, why don't you try an alternitive avenue?  There is a Open Source Org. that works closely with ATI to develope Linux Drivers for the ATI Card.  Why don't you explain the situation to them and beg for a possible fix.  They could make the fix and donate it to ATI; thus giving ATI no negative reason to deny the resolve in the Windows Drivers.  In addition, if the ATI OSO accepts fixing the problem, submitting a proffesional request to accept installing the code into the next ATI driver update would only help marketing value in ATI's cause, thus increasing your chance to getting a sucessful fix.

Besides, if there needs to be a comprimise with the ATI OSO, I'm sure it would be related to getting FSO 100% compatible with Linux.  This in itself is a good thing, because people with Windows could use Cywin to move completely around a possible unexpected problem that may only effect a Windows plateform.  Besides, this move in itself would be awsome for the Linux community.

That's my 2 cents.

[Woolie Wool]

Originally posted by Overlord
Too bad you guys have no Geforce-Card, uh?

nVidia is turning into the new 3Dfx, only it won't die.

[Ace Pace]

The problem occurs in all Radeon cards based on the R300 (I.E 9xx series and the Xxxx series).

So there is no reason to get a mention for each card.

I like xenthorious's idea, how can we try it?

[Turambar]

Originally posted by Overlord
No need, cause it runs pretty fine on my comp.

Runs great on my comp too, but i know it can run better, shadows slow down on the high poly fenris and yet i get about 45 fps in FarCry on highest settings

[phreak]

Originally posted by Ace Pace
I like xenthorious's idea, how can we try it?

i don't know if OSO would like to be recommending Direct3D changes to ATI

[xenthorious]

Originally posted by PhReAk


i don't know if OSO would like to be recommending Direct3D changes to ATI

The point is, if ATI OSO can find the problem, they saved half the problem for ATI, thus reducing cause.

[deep_eyes]

my shine maps work... and im on a 7500.....

[xenthorious]

By the way, PhReAk is absolutely correct.  But that is because I forgot to add a vital piece of information.

ATI OSO focuses on OpenGL support. Hence, what Linux uses.  DirectX support may not be there strong point.  However, Transgamers Cedora team focuses on a DirectX translation into OpenGL code.  They would have experience with DirectX.  In order for Transgaming.com to have a working copy of WineX (now Cedora) for ATI uses, Transgaming.com has to work closely with ATI OSO.  This also applies to WineHQ, who focuses on the wine program.

Now, if FSO was to work with Transgaming/WineHQ and ATI OSO, they could create a fix in Linux.  Since the code will be done in C++, little to no changes would be needed, and would be fully portable over to a Windows platform.  This code would then be submitted to ATI.  ATI would do the necessary translations to incorporate the donated code into there driver.  Thus, a Windows fix is done outside of ATI.  

Now, if ATI does not add the code, then Transgaming.com would incorporate the code, thus making FSO more scalable for a Linux OS.  Eventually, if this path is followed, Windows users could use Cygwin or a duel-boot Linux OS to play the game.  With FRED_OPEN working on Linux, this would provide Transgaming.com an accelerated support for editors , such as Neverwinter Night's Aurora Toolset, Warcraft's Mission editor, etc.  The path would open a huge set of new positive possibilities providing benefits to multiple open source projects.  The close involvement with Transgaming.com/WineHQ could also provide possible additional improvements to FSO.  Transgaming.com could also see a valued beneficiary by using FSO as a a method for troubleshooting serious problems with DirectX9 support with Cedora.

Short and simple, there is a lot of benefits and  possibilities through this avenue.  It could not only provide a fix for ATI drivers, but provide fixes for multiple areas of the Open Source Community as a whole.

[Bobboau]

Originally posted by Turambar
too bad bobbau wont write any nvidia optimization thingys into any code

... ok, like what the FUCK do you want me to do?
what "optomiseations" are there that I have failed to make, and if there were any how in the name of _hell_ am I suposed to test them!?!

[Turnsky]

yeesh.. leave the man be, nVidia, ATI, to be quite honest, i don't give a ****...

Long and short of it..

SCP guys don't get paid for this..
they do it 'cause THEY want to.. and having people constantly pick things, without offering to slightest bit of help.

Turambar, i'm saying this to you... if you want to pick on bob's "supposed" unwillingnes  to write optimisation things (which, i don't see how he couldn't, he and the rest of the SCP team have done a supurb job thus far)
then offer ways that you could help, by actually offering code..
like i said, the SCP don't make FSO for /you/ and you only, but for the freespace community as a whole.

[IceFire]

Originally posted by Bobboau


... ok, like what the FUCK do you want me to do?
what "optomiseations" are there that I have failed to make, and if there were any how in the name of _hell_ am I suposed to test them!?!

Correct me if I'm wrong but I think he's saying that you aren't going to be like some game developers or nVidia themselves and write image quality degrating "optimizations" into the game code for the purposes of boosting scores and getting more people to buy one card instead of the other.

RE: 3DMark 2003 snafu regarding driver optimizations by nVidia (and to a lessert extent and somewhat by accident ATI).

I think that was the direction that was going in...and that you were being complemented for integrity of character

Anyways, I'd love to see this issue resolved.  Cat 4.7 is where I need to be to play my other games.  FSO works just fine but no shinemaps indeed.

[Kosh]

I volunteer Taylor to hack the drivers.

[Xelion]

ATI may never fix this bug, and thats assuming its even related to their drivers or if its even a 'bug'. Does anyone know the exact cause besides narrowing it down to the drivers which really doesn't say much at all?

In the meantime the best alternatives are other solutions. Try the Omega Drivers! Try your latest stock drivers too! If you think this is of major importance, help us find a solution, theres no point asking about it, or pimping it, when its done I no doubt know it'll be posted by someone.

Originally posted by Gregster2k
Video Card & Primary Issue:
- ATI Radeon 9800 Pro 128MB: FSO/SCP Shinemapping implementation completely fails on this card with any driver version above Catalyst 4.4

This sounds like a sporadic case, this really doesn't draw any conclusions to anywhere, try the latest Omega driver.