Carmack on Display Lists

i think he was saying that opengl currently cant do high quality model animation with display lists right now so the best choice is vertex arrays...however, geometery acceleration needs to use display lists to be effective...so when geom accl becomes the standard the opengl api will need to be fixed so you can do model animation with display lists...

#16, like what should we do with all the extra triangle throughput we\'re not using?

From the sound of the last part of the .plan, the solution is not to use display lists for player models... What the alternative is I don\'t know. :)

Maybe that\'s why these GeForce scores suck, because the models aren\'t really accelerated in OpenGL. We\'ll find out when the final GeForce drivers ship and T&L is officially enabled in Q3. But I thought it was enabled all along, I don\'t get this crap about the GeForce benchmarks being skewed because there was no T&L... anyhoo... I\'m dizzy from all the Quake3 news, I have to take my pills.

I donÂ’t *completely* understand the plan update, but i think you\'re right, I think he\'s saying that there IS a problem with accelerating player models in OpenGL. OpenGL allows you to create Display List\'s, which if i remember correctly is a hierarchical list of vertices. This hierarchical structure allows you to translate/rotate (move) one polygon and other polygons attached to that one would translate/rotate as a result. Carmack said that the standard OpenGL implimentation of Display Lists dosent work well for \" high quality skinned characters\" which i belive means that the Transformation will have to be done in Software. Unless NVidia or some someone else comes up with an OpenGL extension that Carmack Likes / works well for accelerating high quality lists of textured and animated polygons. This doesnÂ’t look good for hardware transform acceleration under OpenGL. Well it sounds like NVidia is working towards solving this issue. The question is, did Microsoft solve this problem with DirectX7 if they did, they may actually have some performance advantage over OpenGL. As much as I like Microsoft, OpenGL is simply a lot more programmer friendly than DirectX. So letÂ’s hope we are misunderstanding this issue, or that it is solved quickly.

lol @ #8
...people crack me up.

Seeing that both D3D and OGL support in UT are still pretty bad I think
Tim Sweeny is very very happy with Carmack\'s .plan updates...he\'s getting worldclass advice for free! :)

ON TOPIC POST

Ok, doesn\'t this plan update say that there is difficulty in accelerating player models in OpenGL? It seems to me like only static geometry (levels) can be hardware accelerated. That would not fare well for the GeForce (if it was indeed true), because player models is what causes the big slowdowns in Q3. Does anyone actually fully understand this plan update, and is willing to clarify?

#7 he said late TONIGHT crack fiend

#3... He went to id yesterday night and he said he would give us the info on 1.09 late last night or this morning.... and its past morning!

But who cares when he gets it to us as long as he does sometime soon!

I wonder if the situation would be the same with Direct3D????? ;)

Fine, I\'ll just send the news to Blue...

Aww, c\'mon John. Tell us somehting we didn\'t know already!

9)

-- Martin

So... wheres it at Steve? you know what I\'m talkin\' about :)

Post that email i sent you sTeve.
Im not joking, it looks really cool.