Let me start by saying that only through HAIKU have I realized the truth of OpenGL vs. Direct3D. I had always just thought that because all the game manufacturers supported it most, it must’ve been the best! Lol, hindsight can be 20/20. Im glad to see that HAIKU is based on OpenGL, and now, I hope to bring to the table new light for us in the form of the OpenRT platform; Take alook at this “…fully armed and operational battle-station!” ~Palpatine
I want to express truth now; I think this could be the very thing that draws us up and away from the mundane, am I wrong? Could this not be the very base of HAIKU games, content-creation, indeed the very core of all visio-replecant behavior on the HAIKU platform?
if you’re into that sort of thing, here’s an exchange of arguments by professor Slusallek of the university of Saarbrücken (who advocates hardware accellerated realtime raytracing as the solution of the future) and, on the other side, David Kirk, chief scientist of Nvidia (naturally advocating their polygonal rasterizers).
It’s on a German gaming magazine site, but in English language. Click on “Weiter” on the bottom of the page to browse through all eight pages.
This is not very relevant for Haiku in its current state, though. Even if a complete and stable R1 of Haiku was already released, hardware accellerated OpenGL drivers would be far more crucial to its success than some (at the moment) obscure real time raytracing project.
hardware accellerated OpenGL drivers would be far more crucial to its success than some (at the moment) obscure real time raytracing project.
… unless BeOS were to kick ass at it and be the best performer out there.
Call it a hunch, but I suspect OpenRT would require some type of hardware acceleration in order to be “kick ass” - I haven’t reviewed the site much, but I would suspect they’re making use of exisint hardware 3d acceleration on existing video hardware - possibly using OpenGL acceleration api’s…
"I suspect OpenRT would require some type of hardware acceleration in order to be “kick ass”
Yes, In fact I must apologize for my brevity in exploring thier proposal. Upon further exploration I read that they’re using a newly developed gfx card of thier own, that they’ve dubbed SaarCOR, which is dedicated to the OpenRT ray-tracing API.
Also, please excuse my forewardness in assuming this so soonly attainable under Haiku R2! Once I watched the demo clips, wherin they display the OpenRT platform performing its promises, I saw the text sayng “24 processors, 32 processors, 48 processors:” massively parellel!
Lol, I doubt any of us HAIKU hopefuls will be bouncing around on a 48 processor monster anytime in the soon-far future.
But, let me not destroy the message; The OpenRT project IS aimed at bringing Real-Time raytracing to TODAYS machines, possibly with the help of thier SaarCOR architecture, which appears to consist of very inexpensive components.
And since I had mentioned SMP, in the coming years when GHz chip speeds begin to top-out on phisical limitations, where wil computing likely move? : parellel processing.
I personally will build a Dual Opteron 242 system, a solid reason for my intrest in HAIKU’s parellel processing superiority.
Which begs another question (and another thread somewhere), in considerance of the power of BeOS in regards to multi-treding and muli-processing, will HAIKU become a viable 64-bit platform anytime soon?
“The games where developed using OpenRT, an API similar to OpenGL, which transparently hides the actual implementation of the ray tracer. Like the MESA-System for OpenGL, OpenRT includes a highly optimized software ray tracer, which runs on a single CPU as well as on a cluster of PC (with linear speed-up per CPU).”
…And with clock speeds, memory bandwidths, and bus speeds increasing and memory latencies slowly decresing, in addition to lean-mean fat-grillin’ software optimizations (like HAIKU itself), we could see this abillity coming quickly, no?
“The games where developed using OpenRT, an API similar to OpenGL, which transparently hides the actual implementation of the ray tracer. Like the MESA-System for OpenGL, OpenRT includes a highly optimized software ray tracer, which runs on a single CPU as well as on a cluster of PC (with linear speed-up per CPU).”
…And with clock speeds, memory bandwidths, and bus speeds increasing and memory latencies slowly decresing, in addition to lean-mean fat-grillin’ software optimizations (like HAIKU itself), we could see this abillity coming quickly, no?
I’m sure it sure would be nice if the people over there at OpenRT decided that they wanted Haiku to the be the OS-of-choice to run this platform on…
That might give the Haiku development effort a serious boost as it would have an immediate commercially viable reason to exist
If you wanted to make Haiku a gaming platform, there’s a lot of things you’d need to do… First, get nVidia and ATI to release reference drivers continuously for their cards on the Haiku platform.
Second, prove to some gaming companies that the optimizations within the OS actually do provide a faster working / rendering environment for gaming and physics engines to run.
Third, convince those same companies that there are enough people who would shell out the $0 to get Haiku on their computers and pay the next $50 for their sweet new video game which is going to be released on all platforms, but is suggested to run on Haiku. Blizzard is now an awesome example of a software developer jumping onto a “new” market before everyone else catches on. With the new Apple comps actually using things like hardware accellerated graphics cards and decently quick (games are not photoshop) processors, Blizzard is ahead of the curve in developing and maintaining all of their major titles on both OS X and win xp. As a matter of fact, their new game, probably the biggest one they’ve ever created (http://worldofwarcraft.com), is already available for Mac users to beta test right along with the windows users… Pretty cool.
There’s a lot of games and engines that run on multiple platforms… The problem is getting developers to say “ok, yea, that’s worth develping on or even porting our existing work to”. You have to convince them that there’s money involved, but usually they look at people who don’t pay for an operating system, and they wonder why those people would pay for any software. Hence the paradox linux is just barely starting to bend.
Oh, also, you need hardware developers for other things too to have drivers… IE: Logitech for mice, some USB keyboard companies, and audio companies like Sennheiser and Plantronics who recently moved into having usb-based soundcards built into headphones rather than using your computer’s existing (or not) soundcard to process sound info.
:: Gamer mode off ::
… until I feel the urge to go beta test some more …
If you wanted to make Haiku a gaming platform, there’s a lot of things you’d need to do… First, get nVidia and ATI to release reference drivers continuously for their cards on the Haiku platform.
Second, prove to some gaming companies that the optimizations within the OS actually do provide a faster working / rendering environment for gaming and physics engines to run.
Third, convince those same companies that there are enough people who would shell out the $0 to get Haiku on their computers and pay the next $50 for their sweet new video game which is going to be released on all platforms, but is suggested to run on Haiku. Blizzard is now an awesome example of a software developer jumping onto a “new” market before everyone else catches on. With the new Apple comps actually using things like hardware accellerated graphics cards and decently quick (games are not photoshop) processors, Blizzard is ahead of the curve in developing and maintaining all of their major titles on both OS X and win xp. As a matter of fact, their new game, probably the biggest one they’ve ever created (http://worldofwarcraft.com), is already available for Mac users to beta test right along with the windows users… Pretty cool.
There’s a lot of games and engines that run on multiple platforms… The problem is getting developers to say “ok, yea, that’s worth develping on or even porting our existing work to”. You have to convince them that there’s money involved, but usually they look at people who don’t pay for an operating system, and they wonder why those people would pay for any software. Hence the paradox linux is just barely starting to bend.
Oh, also, you need hardware developers for other things too to have drivers… IE: Logitech for mice, some USB keyboard companies, and audio companies like Sennheiser and Plantronics who recently moved into having usb-based soundcards built into headphones rather than using your computer’s existing (or not) soundcard to process sound info.
:: Gamer mode off ::
… until I feel the urge to go beta test some more …
I was specifically referring to the OpenRT crew (real-time ray tracing) - which isn’t specifically gaming-oriented, although may be useful in gaming. Since they produce their own video hardware to drive their ray tracing technology, the video driver issue is non-existent… they simply need to write a Haiku driver for their hardware. If Haiku provides a solid, well performing platform for their video hardware and ray-tracing software, and it performs better than any other system out there, then it’s immediately a commercially-viable platform, nothing else needed… a high-performance ray-tracing platform sells itself to a commercial market, not consumers.
Haiku has a LONG way to go before it’s a premier gaming OS…
I don’t think Haiku is the perfect choice for running 48-CPU based machines/clusters in fact, I’m sure OpenRT wouldn’t switch to it.
In response to kurtis:
I think Haiku will be able to convince companies that there are people using a free OS who are willing to pay for software/games/etc - BeOS has a good history of commercial software. Haiku is not aimed at the geek “all software must be GPL, easily hackable, and distributed only in source form”, but hopefully at a much wider range of general desktop PC users.
As far as hardware support, I still feel strongly that Haiku Inc could sell an official CD of R1 (at a moderate but hugely-marked-up price ($30-$40 or something)) and from the proceeds have a nice little pile of cash to throw in nVidia’s and ATi’s direction to get some proper hardware accelerated 3D support.
