PDA

View Full Version : catch up with DirectX please!!!



tai4ji2x
05-18-2000, 08:33 AM
I'm worried... to this simple-minded end user, OpenGL seems to be falling behind DirectX... http://www.opengl.org/discussion_boards/ubb/frown.gif

tai4ji2x
05-18-2000, 08:46 AM
I mean... the whole fiasco with Epic announcing that DirectX will be their focus in their future games... (supposedly this is "resolved" now, but I still have my doubts)... this doesn't look good.

john
05-18-2000, 03:31 PM
so one company looks at DirectX, and suddenly OpenGL is doomed?!
There is more to the opengl market than just PC consumer games, you know.

tai4ji2x
06-02-2000, 01:54 PM
yes, but i don't think you can tell me that games and entertainment aren't often the primary "push-the-envelope" driving force in computer technology... or maybe you can? i dunno...

Shayde
06-03-2000, 09:16 PM
Fear not. When you get games like MDK2 coming out featuring breathtaking graphics and effects and actively supporting only OpenGL, there is hope. http://www.opengl.org/discussion_boards/ubb/smile.gif

Epic is only one developer. I suspect their primary motive for preferring Direct3D is the better support for manipulating texture surfaces.

Ciao...
SHAYDE

john
06-04-2000, 03:51 PM
That's right: game's _aren't_ the only motivating factor in developing graphics hardware. People seem to get very excited when the Geforce arrived with transform and lighting engine in silicon. It seems amazing that some voodoo has a hardware accumulation buffer. It is astonishing that all the consumer level graphics cards have all these cool hardware features seemingly to acclerate games.
But! Silicon Graphics/SGI machines have had these things for a lot longer than that. Can SGI justify a multimillion dollar Onyx2 RealtityEngine for games? Do their promotions advocate them as the ultimate Quake machine?
I conceed that, yes, some IR's are used for games, but there is more behind graphics *than* games, particularly on the high end graphics machines. SGI's had this stuff before nvidia crammed it onto a chip for the mass consumer market. But SGIs are used for scientific visualisaion (the visible human, molecules and chemical experiments, volumetric data sets, visualising the mars terrain etc.); for (military) simulators (where theres loadsamoney from the USA goverment); visualising nuclear explosions; rendering visual effects for films; and so on and so on.
It may well be games that push this technology onto the little computers, but it isn't necessarily games that push the technology on the big machines.

my 2c worth.
cheers
John

tai4ji2x
06-04-2000, 05:18 PM
sorry about the confusion everyone. yeah, i'm mainly talking about us "little computers" http://www.opengl.org/discussion_boards/ubb/wink.gif i came to this forum just as an end user, not an opengl developer or graphics professional. so i' a little game-centric. http://www.opengl.org/discussion_boards/ubb/tongue.gif sorry. thanks for all the input though.

Mike F
06-05-2000, 03:24 PM
Originally posted by john:
That's right: game's _aren't_ the only motivating factor in developing graphics hardware. People seem to get very excited when the Geforce arrived with transform and lighting engine in silicon. It seems amazing that some voodoo has a hardware accumulation buffer. It is astonishing that all the consumer level graphics cards have all these cool hardware features seemingly to acclerate games.
But! Silicon Graphics/SGI machines have had these things for a lot longer than that. Can SGI justify a multimillion dollar Onyx2 RealtityEngine for games? Do their promotions advocate them as the ultimate Quake machine?
I conceed that, yes, some IR's are used for games, but there is more behind graphics *than* games, particularly on the high end graphics machines. SGI's had this stuff before nvidia crammed it onto a chip for the mass consumer market. But SGIs are used for scientific visualisaion (the visible human, molecules and chemical experiments, volumetric data sets, visualising the mars terrain etc.); for (military) simulators (where theres loadsamoney from the USA goverment); visualising nuclear explosions; rendering visual effects for films; and so on and so on.
It may well be games that push this technology onto the little computers, but it isn't necessarily games that push the technology on the big machines.

my 2c worth.
cheers
John

Take a look at the extensions registry at http://oss.sgi.com/projects/ogl-sample/registry/

The extensions that provide major new functionality have recently come mainly from companies like nVidia, ATI and 3dfx, all looking to differentiate their products for gamers.

Mike F

john
06-05-2000, 03:38 PM
yes, I conceed that games are a factor in extending opengl. My assertion was that they are not the *only* factor. Keep in mind that the initial opengl specification was based on silicon graphics' API's, so already a huge proportion of opengl has come from a non-game sector.

pop_n_fresh
06-11-2000, 09:50 PM
I'm curious, what specifically does Direct3D have that OpenGL doesn't? Because of the extension mechanism new hardware features are often available in OpenGL before Direct3D. Multitexture and Register Combiners are two examples. NVidia's Register Combiners functionality will not be available until DX8 I believe. As for Epic, the problem isn't so much with OpenGL but rather the Windows driver model which is why they're dropping OpenGL support for windows. Microsoft, of course, in rather unmotivated to fix the problem.

synergy
06-14-2000, 03:06 AM
First of all, I agree with "john" mostly, but not on the last statement. What are you talking about "little computers", "big computers". What's pushing the development in computer graphics is the great computer scientists & mathematicians all around the world. Algorithms get developed and refined through time. Some with more- others with less sucess. Most of the algorithms ware developed in the 70's & 80's by the great pioneers of computer graphics to finally end up in some API like OpenGL. What we are talking about here is that OpenGL needs to get better as time goes by. Nothing remains on top without reviding it. Of course OpenGL is a nr 1 choice for 3d data analysis in research laboratories/military, but it would be a shame if OpenGL had to draw back from the enertainment industry due to lack of refreshment of its standardization.

Secondly for people who think extensions solve the whole problem...

Extensions are great but more vendors need to support them, not only like NVIDIA. Moreover the entertainment industry must support more extensions in the future. What's the point of adding a lot of extensions to OpenGL if only a couple of vendors & developers support them. Add them to OpenGL itself and make them standard.

I think we really need OpenGL1.3/2.0 with big improvements not like from 1.1 to 1.2 (almost no improvement).

One should never underestimate the power of Microsoft. Many people say that they are unable of making DirectX as popular as OpenGL, but think a while... They got money and lots if it. So they can hire experts in computer graphics. Already many great computer scientists work for Microsoft. Some of them are worldleading in their reserch topics.

OpenGL has an advantage having multi-platform support. Unfortunatelly that's not a big advantage, OpenGL needs to get better, and really fast indeed.

You may think I'm a big fan of Microsoft, NO. I was anti-Microsoft, til I realized, Hey they are actully producing quality software nowdays!

I don't care who's taking my money, I only care for quality products.

I'm using OpenGL for 3 years now and seen actually no major improvement. Sad to say it, but that's the fact. I think many of us wants big improvements on the standardization of the fortcoming OpenGL version.

Otherwise OpenGL as a 3D game API is probably doomed...



[This message has been edited by synergy (edited 06-14-2000).]

ngill
07-01-2000, 09:07 PM
Originally posted by synergy:
I think we really need OpenGL1.3/2.0 with big improvements not like from 1.1 to 1.2 (almost no improvement).

One should never underestimate the power of Microsoft. Many people say that they are unable of making DirectX as popular as OpenGL, but think a while... They got money and lots if it. So they can hire experts in computer graphics. Already many great computer scientists work for Microsoft. Some of them are worldleading in their reserch topics.

OpenGL has an advantage having multi-platform support. Unfortunatelly that's not a big advantage, OpenGL needs to get better, and really fast indeed.


I know dude, I don't want to spend my years masterning OpenGL just to find out ten years down the road it is obsolete!! Microsoft is a lazy bum! They do not make good products as in COMPATIBLE!

But, I wish OpenGL was like really ahead of DirectX, so that video card manufacturers would put OpenGL ICDs before support direct junk. That would be a big inspiration...

Damnfuct
07-17-2000, 03:56 PM
Personally, i find that OpenGL doesn't just look better, but it also runs faster. I own a geforce, and i find that it runs OpenGL faster than directx, and wherever its possible, i use opengl over directx. microsoft makes directx, enough said http://www.opengl.org/discussion_boards/ubb/wink.gif

subpoposaur
07-19-2000, 08:03 PM
I don't know how interested you all are in this, but there is a largish project involving just about everyone in 3d hardware including sgi and intel called openML. This is, from what I have read, intended to be the openGL style equivalent of Dsound etc.

Hopefully this will give us an all in one api for graphics/MM/games/pretty business apps??
http://www.khronos.org

[This message has been edited by subpoposaur (edited 07-19-2000).]

Kilam Malik
07-21-2000, 07:55 AM
Nice descriptions on the OpenML page, but how will they compete with OpenGL and DirectX? There are years of development and the company which has reached the most influence in graphic boards market, NVIDIA, is not on the list of supporting companies. So I don't give OpenML great chances.

To the speed issues: I tested DirectX vs. OpenGL and DirectX runs about 1% faster. So I think the speed is nearly the same, OpenGL is not faster. And how could it be? DirectX is more integrated in the system.

Regards,

Kilam.

Whittick
07-21-2000, 08:23 AM
Speed of either api is dependasnt on the graphics board and the drivers for it. Take a Geforce for exaple the lateset drviers (5.22) had heavy T&L opengl optimizations and hence you can see the performace difference. Tho in my personal experience (on a Wintel) D3D is usually a bit faster but I have noticed OpenGL looks a lot nicer.

For an example of this play half-life in Directx and then in OpenGL and you will see what im talkig about.

Chris

Whittick
07-21-2000, 08:31 AM
Kilam im not flaming you here but NVIDIA might be a big name but you make it sound like these are nobodies:

3dfx, 3Dlabs, ATI, Compaq, Discreet, Evans & Sutherland, IBM, Intel, S3, and SGI

Nvidia is a major player in the consumer level(destop) market and they are starting to move more and more into the workstation area but look at the list. All those are major major players in the computer industry itself and all they are going against basically is microsoft (i know i make it sound llike a small thing). So i dont see it failing at all. I hope it introduces a great programming API for years to come.

Chris

MikeC
07-21-2000, 10:45 AM
I know I've said this before, but a lot of people seem to have a false impression of what OpenML is.

OpenML is not a better OpenGL. It's not a cross-platform DirectX. It's aimed more at high-bandwidth streaming media (audio and video). The main focus looks like being on combining multiple streams, e.g. resolving timing and synchronization issues.

OpenML will (optionally) run in parallel with OpenGL. They mention that some OpenGL extensions are planned, which will probably have to do with sharing the video framebuffer (e.g. allowing hardware 3D and hardware MPEG in the same frame without a load of cruddy GDI blits or glTexSubImage workarounds.

I was surprised that NVidia wasn't on the list - these days they're normally in sync with whatever SGI is doing. Cass, anything you could tell us without having to shoot us afterwards?

ngill
07-22-2000, 07:35 PM
[War3ChatHost] Why the decision not to have OpenGL as a 3D acceleration choice?

[War3DevTeam] Direct3D is a much better choice for our game because we do skinned, skeletal animation on our units. OpenGL in its current form really can't support hardware transformation and lighting on those types of units.

HasrihnTokini
08-03-2000, 10:59 PM
SGI hasn't released the source to their implementation of OpenGL have they?
I don't think so... which pisses me off in a way...
If it were open sourced and had one official site where new versions are compiled, then it would progress much better, as everyone who had a mind to would go make modifications to the API and submit these modifications, and if they were good, usefull, and clean, then they could be added to it...

just my two cents...

--the Rabid_Stick

Mike F
08-04-2000, 05:35 PM
Originally posted by HasrihnTokini:
SGI hasn't released the source to their implementation of OpenGL have they?
I don't think so... which pisses me off in a way...

1) SGI has in fact released the entire source to their OpenGL sample implementation (plus GLU and GLUT
2) Even if they hadn't, there's a 3rd-party open-source implementation of OpenGL available (Mesa).
3) Even if neither of the above were available, that's no reason to be pissed off; SGI is a commercial company and they don't owe you anything.


If it were open sourced and had one official site where new versions are compiled, then it would progress much better, as everyone who had a mind to would go make modifications to the API and submit these modifications, and if they were good, usefull, and clean, then they could be added to it...

Nope, hasn't happened yet. Although it has been frequently mentioned as a useful way of evaluating and generating interest in new features, it seems that the trouble involved in downloading Mesa or the SI, getting it to build and then implementing a mod is too much trouble.

Mike F

kwutzke
08-13-2000, 04:03 PM
Integrated into Microsoft. And anytime they realize their design is no good, they do it all over and you have to learn their new mistakes again, and again, and again, ...

...Though OpenGL really has to move a little faster. Otherwise the OpenGL community will end up using large amounts of extensions, which will make the software "incompatible" again. This is not what OpenGL is supposed to be used for. Maybe picking up some more useful extensions into the spec would "unite" the OpenGL hardware/driver manufacturers and will make development less restrictive and unnecessarily complicated.

Just some thoughts.

vs
08-16-2000, 07:21 AM
I am an ex directx programmer. I work doing 3D development for a retail organization. What we did'nt like about directx is that its changing all the time and not necessarily upwardly compatable. It does not function properly under windows NT and has a specification which does not need be supported by HEL or HAL to be certified directx compliant????? Pathetic. To us OPENGL is fantastic. Yes there may be a little more leg work required in certain areas to get the desired result, but the result is better. The platform is stable. Its more flexible and implements into MS windows cleaner. There are great books and manuals describing it in detail. The biggest bonus however is that I don't have to talk to the MS directx bozo's who will say anything they think that you want to hear. You would think they were running in an election.

My only concern with OPENGL is :

1 - that it will make a knee jerk reaction and add to much functionality to fast

2 - in a multi viewed app there seems to be no way of determining if another application is running and using a particular ICD with the graphics accelerator, which will cause any view becoming active that uses the same ICD to hang.


Smile, be glad OPENGL exists and that you are using it and not DirectX. Alternatively switch to directx, be depressed, and hope america does not elect another member of the bush clan.

skippyj777
08-17-2000, 02:48 PM
As far as speed is concerned, I have a PII 350 with 320 MB of RAM and a Matrox Millenium G200 8 MB AGP card with the lastest w2k drivers (I use w2k).

When I played the Descent III demo in DirectX mode, my settings were 640x480x16 with a 16-bit zbuffer. The FPS was around 25. When I ran in OpenGL mode (which the game did not recommend), my settings were 1280x1024x32 with a 32-bit zbuffer. The FPS stayed between 20 and 40. Image quality was slightly improved in OpenGL mode.

Now, if I'm not mistaken, doesn't that mean OpenGL is just a *little* bit faster than DirectX 7 on my system? http://www.opengl.org/discussion_boards/ubb/smile.gif

dorbie
08-25-2000, 06:34 PM
I'm curious,

In what areas would you like OpenGL to catch up.

Other than vague or misleading nonsense in the press, OpenGL is not actually behind as far as I can see. So, please what specific features or in what areas do you see OpenGL as lagging behind DirectX?

skippyj777
08-27-2000, 01:03 PM
I noticed one feature lacking, but only one: no support for rendering directly to a memory pointer (neither system nor video memory). I would prefer to completely bypass the Windows GDI bitmap functions and simply allocate a block of memory and use a single function call to tell OpenGL to render to it. DirectX DirectDrawSurface objects are basically blocks of video or system memory that can be read to or written to directly. Direct3D renders directly to these memory blocks, so you don't have to mess with the GDI.