Unigine Heaven On Linux In A Month Or Two

[L33F3R]

dragengine?

[mtippett]

The older gpu tesselation support isn't nearly as powerful or flexible, it's essentially what ATI developed for the XBox360 GPU. It was never used on PCs because NVidia didn't have anything comparable and developers apparently decided it wasn't worth using if only half their customers would benefit.

It goes beyond that too. Before DX11 was defined, tesselation was a neat idea, but remained an extension that was ATI specific.

Now that it is included in DX11, suddenly it is a lot more important - it will allow PC gaming jump up in visual fidelity quite a large amount (without causing large costs in non-GPU model complexity. All mesh changes and transforms are done on the simpler model (6k triangles for the dragon in heaven), but are then bumped up to the millions of triangles on the GPU for render.

I am not sure if the OpenGL tesselation extension will reach back into older GPUs, but there may be HW specifics that either the extension needs, or that Unigine relies on.

Regards,

Matthew

[wien]

"Unified shaders" were introduced with promise that they can be programmed to do generic work. Now they tell us they actually can't and to get new hardware instead.

It's always like that. First make promises, then break them.

Maybe if, instead of whining on internet forums, you'd actually educate yourself slightly about graphics hardware, you wouldn't end up so disappointed. You're making up stories based on what can only be described as peripheral knowledge and hearsay, and then holding it against IHVs when things don't turn out the way you fooled yourself into believing they would.

[Dragonlord]

Unfortunately he's right. That tessellation stuff is another "holy-grail" of the graphics industries which doesn't hold up to what it promises. Not the first time this holy-grail syndrome happened and for sure not the last time. I would definitely prefer a game/engine doing proper lighting and what else to one which wastes hardware on something like tessellation which has a good deal of problems attached to it. It's anyways a bit questionable to waste hardware on tessellation but on the other hand you have no shadows at all for anything else but the sun-light. If something is not tessellated ( even assuming you fail at doing normal mapping properly ) it's less of a visual problem than missing shadows or other interesting gadgets. And locking yourself in with DX11 for that is another questionable move.

That said though it's their decision. Not much for me to question in the end so I guess I'll go back to the non-speak zone ( and anyways didn't wanted to post but sometimes you just can't sit on your mouth any longer :P )

[mtippett]

Unfortunately he's right. That tessellation stuff is another "holy-grail" of the graphics industries which doesn't hold up to what it promises. Not the first time this holy-grail syndrome happened and for sure not the last time. I would definitely prefer a game/engine doing proper lighting and what else to one which wastes hardware on something like tessellation which has a good deal of problems attached to it. It's anyways a bit questionable to waste hardware on tessellation but on the other hand you have no shadows at all for anything else but the sun-light. If something is not tessellated ( even assuming you fail at doing normal mapping properly ) it's less of a visual problem than missing shadows or other interesting gadgets. And locking yourself in with DX11 for that is another questionable move.

There are some nice benefits of tesselation.

1. You deal with a simpler model in your engine. Collision, transforms, movement, etc are all done on a simple < 10k model.
2. You need to carry in a lot of cases less content on the host side, particularly with procedural tesselation
3. You push *way* less geometry around. The GPU renders millions of triangles, your engine pushes 100's.
4. If you have shadows (like Unigine does), your shader based light model can interact with the actual tesselated geometry.

If you look at the shadows from the dragon with tesselation, the bumps and so on are also included in the shadows. If you don't use tesselation, you'll just get the simple shadow.

Obviously if you do host side pre-rendering, you don't get the tesselated geometry, which may be result in issues that you highlighted. But I believe that Unigine's light model is almost entirely shader based.

[Duo Maxwell]

Boohoo, there are people out there who like deathmatch and there are a lot of linux games out there with deathmatch...

Sorry guys but your getting repetitive and cheap. I understand it must be frustrating that there aren't many games of your favorite genre on Linux but stop beating the dead horse please. We have heard the complaint over and over now here on Phoronix and we are aware of it. Be happy for the deathmatch lovers for a change.

We need another run and gun fps title on linux like we need another hole in our heads, I dunno about you but aside from a match or 2 every few weeks in QL is way more then enough FPS. I'm bored to death with quake remakes on linux and nobody is even making an attempt at making a single player fps game like the S.T.A.L.K.E.R. series that I got around to playing only becaue I borrowed a friends windows box for a few days.

If linux gaming is going to be taken seriously it needs to break out from the "id just released the idTechover9000! engine, lets make yet another uncreative, repetitive Quake or UT clone that almost nobody will play regardless that it's free or not because by the time we have a working title everybody and their cat has a copy of UT 2k25 that they picked up for $2 at Crazy Sal's Bargain Hut."

If you're going to make yet another fps you could at least make an attempt at not doing the same as every other linux fps, the same game just with different models and varying levels of graphics polish.

Don't give that crap line about "nobody will give art for free" theres tons out there, google around, whatever else you need start bumming around forI'm sure theres a few students out there.

[wien]

That tessellation stuff is another "holy-grail" of the graphics industries which doesn't hold up to what it promises.

The only people calling it a holy grail are marketing folks and people who don't know what they're talking about. This will eventually lead to people whining that it was in fact not the holy grail they thought it was, which was my point above. It's the same every goddamn hardware generation, be it GPUs or consoles, and it's starting to get real old.

For developers tessellation is just another tool for the toolbox. Something they can use to do some cool stuff. Nothing more, nothing less.

[Duo Maxwell]

NVidia looks like the untesselated version, I believe they want to wait until the demo can be viewed "as intended" - with tesselation.

Do I detect a litle *poke* *poke* *nudge* *nudge* ?

[Kano]

Well ATI seems to shine now until Nvidia has got DX11 chips... But the driver does not seem to be good enough for the demo, oh who did expect that Even John Carmack sees problems with fglrx and he is much more important than Unigine...

[L33F3R]

Don't give that crap line about "nobody will give art for free" theres tons out there, google around, whatever else you need start bumming around forI'm sure theres a few students out there.

I dont know if its my scene or just that people dont look... but i know plenty of talented artists that only want acknowledgement for the work they do. Im not talking about moddb talent either . This can best be described as pretending to be on a team for 2 weeks and then disappearing.

Artists just want to be loved. It should be no surprise that they dont like FOSS projects based on IDtech X because no one plays then. They are old, lack creativity, and are all generally the same. No one wants to do art for a game that no one is going to play....

Nexuiz and alien arena, bend that bar a little bit. But at the end of the day its mostly the same stuff. Take open arena: We can all agree that it is ugly as hell. I have offered my services to the team previously; but they seem content on having dirty ass maps with 2 walls, a floor and little else. When a dev team does this, the blame is placed on them, not the world of artwork.

I will be the first to know that programmers will build a game around what they want, and we have to live with that. If we want to mix it up a bit then we need to get creative. Until then its deathmatch on linux.

And may the double damage be with you.