Global Illumination suxx nature!

To Lauscher: There are some problems with our download server.So i redirected the link to another one. Now it should be running. greets, bull / federation against nature

PLEASE support dual-core cpus and make the new benchmark multuthreaded. Would be great if you patch the old realstorm04 benchmark too, so that it supports dual-core cpus. THX.

Hi, i can´t download the bowl x-treme demo, i always get redirect to another page, but the download does not start / i can´t find it. can somebody help me?

Great work guys!! Can't wait to see the new bench! But to answer MKS: There won't be such a thing as a raytrace-unit on gfxboards soon as the existing hardware-RT-prototypes can only handle more or less static objects, whereas every game released nowadays features fully interactive/dynamic scenes (=which is evil for ALL existing ray tracing algorithms)..

Great work as always. Please make the new benchmark multi threaded so it can be used in SMP / Multi-Core benchmarking. Best Wishes.. L.

Wow, very good Demos, and thy run with a normal PC with an good CPU. I hope that in near time every Graficard has a RaytraceUnit to calculate raytracing.

Herorragende Arbeit! Gratulation! Und vielen Dank für die tolle Arbeit. Keep up the good work! Und falls Ihr Zeit habt wäre eine Linux-Version toll. RAYTRACING ist die Zukunft!

I like a lot your demos and Realstorm. Why not to use SDL (www.libsdl.org) to make the stuff easily portable through platforms? Respect! -Mathieu-

Hi there, Nice engine! What I like most about it is that I do not have MS-Windows, and run it under 'wine' in Linux! Its runs perfect, considering it is running emulated! On my Athon64 3000+ and FX5700 (425MHz) I get around 12-14 fps in the benchmark...not fast but certainly nice! Is it possible to port your engine natively to linux? :-) Kind regards, Chris, Brisbane, Australia.

Very good Demo's, ran nice on my FX-55...Duel processors wont be good for gaming, what do you do?

Love your demos and your work. Can't wait to see what you can do once dual core processors become "mainstream".

hi tts :) schön zu sehen das es dich noch gibt musste heute irgendwie an die c64 zeiten denken und hab mal ein wenig rumgesucht im internet. nun bin ich hier gelandet. hammer was du da wieder fabrizierst *g ein gruß auch an alle alten oxyron member alles gute Thomas . Peace/oxyron Ps: würde mich über ne mail von dir freuen.

Very promising! I hope we will see support for Multicore/SMP and Linux soon.

Very impressiv ! That is optimising ! Please, don't forget that new hardware can be programable with Cg for Real time raytracing.

TTS, ich glaube du drehst langsam durch ;-D. Freue mich auf die nexte Version. Greets and Dich und Bull.

i dey here you guys.......................................

Hey,eure Engine ist super geil. Ich arbeite auch an einer, doch mit Raytracing seit ihr hoch überlegen, SSSUUUPPEERR COOL eure Engine.

Atemberaubend. Breathtaking.

elite

Checkout these articles from gamedev.net. It gives a fast method of simulating radiosity for real time use: http://www.gamedev.net/reference/programming /features/rtradiosity/ Not sure what your current implementation is like but this might help you speed up a few frames.

just amazing.

Raytracing in realtime - really impressive. Since this is a long time project you may have a look at the Shader Model 3.0. With it's dynamic flow control in the pixel shader it would really help making RTRT even better and - much more important - faster.

Wow! Das ist echt cool! Wird man diese Engine für Spiele nutzen können? Das wär schön für euch - und toll für uns ;)

Das ist der Wahnsinn ich such immernoch die *.avi irgendwo verstekct in einem ordner ^^

Impressive routines but I didn't expect to see anything else here. Excellent work, as usual. Thumbs up.

Smooth stuff, keep it up !

Geil Geil Geil. Das was ihr hingekriegt habt ist unglaublich. Macht weiter so!

Very nice Game!!!! THIS IS THE FUTURE!!! CONGRATULATIONS!!!!

@GfX & Darkness: ihr idioten! kniet nieder und zollt den entwicklern den respekt den sie verdienen, ihr kiddies! nochmal: alles was man in den demos sieht ist komplett polygon-frei und wird von der cpu durch mathematik uws berechnet(ist technisch (noch!) unmoeglich mit der graka zu berechnen)! ausserdem sind die schatten korrekt und die spiegelungen realitisch anders als in zB. farcry... @FAN: good job! you rock!

diese scheiße ruckelt ja mehr als farcry auf 1600x1200! Sieht aber ganz gut aus :)

This demo sucks!

Thumbs up for amazing work! I read that the engine is in pure C/C++. No assembly, no explicit SIMD stuff. I understand this is because you want your code 'clean' of vendor specific stuff. Have you consider/tried different compilers that do vectorization themselves? How do thay fare? Future will bring multicore CPUs, so single thread performance is probably not going to increase with such tempo as it has before. Have you considered making the engine multithreaded? I don't know much about the innerworkings but raytracing is supposedly very easy to paralelize - just fire multiple threads chasing their own rays.

Hm... Ever thought about offloading some of the workload onto a modern GPU? Far shot, but maybe the "BrookGPU" Project is worth looking into for you guys : http://graphics.stanford.edu/ projects/brookgpu/ (Project aims to create a Compiler to allow 'normal' Calculations/Code to be run on a GPU instead of the CPU where suitable; obviously has limitations, but certainly has massive potential in the future)

Bin wirklich beeindruckt! Raytracing ist zum einen mal was anderes, als dieses ganze 3D beschleunigte Zeugs... zum anderen sieht es verflucht gut aus!

By the way: The rendering speed of a hyperthreading processor increases to a maximum of about 125%, NOT 200%! One more thing: Your fake radiosity algorithm produces very bad results, especially in the "radio"-scene. There you can see that the radiosity is faked by a far too sparse array of "diffuse reflection" lightsources. Furthermore the array itself is poorly arranged. How about some quick and dirty matrix radiosity as an alternative option? At the moment the artifacts are terrible and the light diffusion covers not the whole solid angle it should (see the rgb-balls). You did a superb job nonetheless! ALBEDO

By the way: The rendering speed of a hyperthreading processor increases to a maximum of about 125%, NOT 200%! One more thing: Your fake radiosity algorithm produces very bad results, especially in the "radio"-scene. There you can see that the radiosity is faked by a far too sparse array of "diffuse reflection" lightsources. Furthermore the array itself is poorly arranged. How about some quick and dirty matrix radiosity as an alternative option? At the moment the artifacts are terrible and the light diffusion covers not the whole solid angle it should (see the rgb-balls). You did a superb job nonetheless! ALBEDO

OK. Updated the database browser. Now I see my Processor as Athlon 64. :-)

Hi! Nice GFX! The Benchmark database declares my Athlon64 just as Athlon processor. maybe you should change it for more readability. I would like to see some photon mapping for caustics or more complicated brdf models. How about some spectral rendering for diffraction or interference phenomena? Maybe you can add hdri and motion blur. etc., etc. ...... You're doing a good job! Happy rendering, ALBEDO (heavy Cinema4D user)

I know you don't use the 3d card, and I know why, but with programmable graphics processors you actually could. You could, for instance, use something like deferred shading (http://www.pvrdev.com/pub /PC/eg/h/DeferredShading.htm) to calculate the lighting, and I don't really know but it could be faster to pass the texture coordinates to the graphics card and let it find the correct pixel. I doubt that though, and you would have to store an entire screen of texture coordinates on top of everything. I don't expect you to do that, you'd probably have to start again, but I think it's interesting that there are now (extremely innefficient I expect) raytracers that run entirely on the graphics card and raytracers that run on the cpu, but no hybrids (that I know of).

you guys really need to add mipmapping. currently all diminished textures are showing severe aliasing artifacts. and yes, mipmapping also applies to raytracing, for those who are wondering.

Great benchmark! :-) Will you add SMP support anytime soon?

this is hm... let's say at least EXTREMLY AWESOME ;)) ok, now where do I download your library ;)?

Incredible but needs support for multiple processors. Half of my HyperThreading Pentium 4's processing power is going to waste!

MUGUOOOOOOO I DEY HEREOOOOOOOOOOOOOOOOO

Incredible work! It would be nice to have a port for Windows XP, and maybe for Linux! Does it support hiperthreading?

I think you need a bit more cpu support in there somewhere... a 2.1ghz athlon with 512mb should not be faster than a dual p4 xeon 3ghz with 2gb!

nice work ^^ cool wär noch wenn ihr n forum dazu einrichten könntet :)

Most impressive, the realtime radiosity and those volumetric effects left me gobsmacked! I am definately a FAN of Realstorm ;)

Hi, Nice demo, but I am missing bilinear filtering on some textures as well as mipmapping. Texture shimmering and moiree artefacts were very ugly, but the rest was brilliant - as expected for a raytrancer ;-)

Good job guys. I commend anything that looks this great and brings my systems down to their knees. Keep up the good work!

It would be really nice to be able to start the benchmark process, leave, and come back the next day with results from all the different configurations. Is there an option in the RealStorm Bench2004 program to run through the (10) most common benchmark configurations and then have them automatically save to the local database file? If not, could this be added for the next version?

I saw someone said that you're engine left no time for physics. That made me think: If you want to have the only realtime raytracer with physics, then check out www.tokamakphysics.com. I tried the demo with 20 ragdoll's all landing on each other and the most time the physics ever took was 8 miliseconds. I only have a 1ghz processor too. I really doubt you'd notice that with how demanding that engine is. And realisticly simulating 20 dead people landing in a heap on some steps is a more demanding task than you're likely to want.

is the angine free for use in freeware gamedev ? i found no information regarding usage of the engine on your website. i would be pleased to hear something from you. look at my homepage: http://www.andre-krause.net . maybe your engine would be the first to render rtrt tetris ?!?

Superb! now i cant truly test my CPU & RAM! thankyou to those responsible for this application, And ppl that use this or are going to then READ the READ ME file within the archive it sates "4. What Hardware is used? ------------------------- The RealStorm Engine calculates everything with the processor and uses a lot of memory, it does NOT use your 3D-Hardware. You may ask why does it not use my 3D-Hardware? The answer is simple, your 3D-Hardware can't do what the RealStorm Engine does." So again i say thankyou for the true calculations regs, ravenmeifter / Independant (Australia)

Extremely cool Think someone else asked about multiprocessor speeding this thing up but I was wondering, can you use multiple machines with a high speed network(gigabit or some such) to get more performance? Anyway, this is going to be the future of 3d visualisation generally and it actually runs at an bearable speed on my humble machine! :)

Good day. How can i get the full version of your engine? We are game developing team and we making a new RPG game. We think we can make intellenget rpg as fallout + very good action. Our project is not commercial, but if we can find publisher we can pay you for your engine. We have very good modellers, animators and other professionals. I would consider that we could have mutually advantageous and very fruitful cooperation. MegagerZ

good work!! have you ever considered to create a hw version of your renderer? (i mean by developing a custom chip or licensing the renderer) practically this is probably not really usuable in commercial software since there is e.g. not much (if any) cpu time left for a physics engine (got ~13fps on my amd barton2500+ which otherwise rocks with hl2 ..) a 3D plugin / viewer for browsers would be nice in order to use this engine for 3D rich media applications (e.g. product portfolios..you probably know the "viewpoint" engine..) anyway.. this is also a very nice prototype for the future generation of 3d cards-- and its cool to finally see something different than all those "standard" h/w 3d routines used by most demos released today keep on hacking ! :)

THIS IS THE BEST BENCH MARK EVER!!!

New benchmark is sweet. Liked the demo #1 scene best, but #5 was a nice touch.

Hi TTS and Bull, good that you released the new version before wintertime, so you make sure that the cpu keeps the room heated well :-D very nice new features and another bunch of stunning new effects... keep up the good work and always keep in mind our old slogan: do it one-framed or don't do it ;-D

Glad to see you guys are still pushing the mark!! :) Hope to see more of you guys in the scene soon again! :)

hehe.. the new bench is great. i have to congratulate you.. now i'm looking for first athlon64 raymark results:D

I am just signing to the guestbook. The RTR engine is simply wow amazing!

hey ppl, didn't see ya for a long time! Any meetings left?

Impressive. I especially like your use of booleans (in the benchmark). Have you considered photon mapping for lighting, instead of radiosity? You seem to have the ray intersections very streamlined, so it wouldn't be too much of a challenge. Did I read that there is no assembly code? SIMD instructions should give you a 4x performance boost. bolts

Hey Michael. I got a name for the next demo: "Suck in the nature" *S*. Nice to see your latest demo. Keep Cool.

Lovely & useful utility, great site ;) Keep going ppl !

Hey Markus Seems you haven´t got any Idea of coding. Sorry i think you are nothing but a lamer.

Yeah, the programms are very impressiv from a programmers POV, but ultimatly they SUCK. The benchmark and the demos look like 97 Demos. Low framerates even with high end systems and the notorious undersampling artifacts arent set off by a few nice reflections that could easily faked with EMBM or pixel shaders. Yeah, it is quite an archivement to get even that much speed and quality out of raytracing, but i cant stand all those people saying "yeah, low res, shitty framerate, but look, its realtime raytracing, how impressive!!!".

nice work

WOW!!! I'm really impressed guys, keep up this cool work!

Have'nt checked out the Scene for a couple of years. Got bored seeing the same old shit. NOW THIS IS NEW! Full Marks Guys, You are REALTIME GODS ! Keep up the good work ( any chance of the source code ? ) ( :

lol @ elvis. the 'no-gpu' idea is what makes this engine so nice. all your nice geforce and radeon cards can only do what they are programmed for. if u ever want to have an effect the card doesn't support, u're busted :|

Nice, but where is the performance? It makes no sense to code an engine which uses the CPU in such an extensive way !!! We have GeForce and Radeon! This engine is beautiful but useless for games and stuff... Try selling it to Discreet... their raytracing in 3D Studio Max is very slow... good job... keep on coding

Nice, but where is the performance? It makes no sense to code an engine which uses the CPU in such an extensive way !!! We have GeForce and Radeon! This engine is beautiful but useless for games and stuff... Try selling it to Discreet... their raytracing in 3D Studio Max is very slow... good job... keep on coding

How likely is it that a linux version will be written? This could be a great cross-platform benchmark and library.

Okay, nice engine - I'm impressed! But why the hell are you against nature??? That is the dumbest advertisment sheme ever. "Huh, we gotta be against something ... let's be against nature!" "Yeah, that sounds cool" "Yeah - huhuhuhuuu"

¿ Only Windows ?

Will it work faster on multiprocessor systems?

Very impressive! Is this engine for sale or download?

Never seen such good raytracing very imprecive, i whould like if someone give me links where i could read more about this technology (math part of this) i started to learn about it not far ago

hmm, da liest man die netten kommentare und was wünscht man sich? genau, dass es auch auf dem eigenen computer läuft. deshalb meine bitte: könntet ihr die demos nicht auch auf freeBSD/MacOSX portieren? naja, ein versuch wars wert ;-)

THIS DEMO KICKS MY COMPUTER'S ASS. 16FPS ON 640X480 WITH ATHLON XP 2100, 512MB 333MHZ RAM, AND RADEON 9700PRO

Im really impressed guys, cool work!

Whoa!!!! Eyes pop open. Cool!!!!

wow.

Rockt mindestens genauso wie Oxyron damals :) Keep Up....!

really nice demo(s) :) just wanted to note that if u manage it to "port" your raytracing engine into a streaming model (on the inet there are lots of infos about this) u could use the power of modern gpus (~50 gflops on a gforcefx or ~200 on a p10(wildcat vp)), u "just" have to code some pixelshaders. imagine a scene 50 times as complex as yours rendered in 1024 with 30fps...

Check http://www.openrt.de for some really impressive RTRT on a cluster

Great work. Is there a chance to get a API?

Hi ! Fantastic work. How did you got it that fast ? How many of your code is written in asm ?

Nice work. Federation Against Nature (and stupid socialist shitheads) !!!

Nice work. I noticed that antialiasing give better average and worst-case results.

I am puzzled. Where do I get the program? I downloaded the only item that appeared to be an application, the latest Benchmark.exe, expecting to get some kind of demo or an engine with a user interface with which one can create rendered images. If this is shareware, where do I download it. If this is for sale, where do I buy it.

(ARGH, the guestbook is erasing html-tags) Hy TTS, your engine is just a hidden feature in windows... look here: http://www.mdc.de/images/TTS-Engine.gif ;) Just kidding... TIN hat bald wieder ne richtige Matte aufm Kopp :)

Hy TTS, your engine is just a hidden feature in windows... look here: ;) Just kidding... TIN hat bald wieder ne richtige Matte aufm Kopp :)

TSCC WANTS "KNIGHTS" ... NOW! :)

Realstorm looks great, sometimes even more real than those fancy T&L hardware-accelerated polyengines.. You absolutely want to tell me some tricks about object rotations and stuff.. :) - Mankeli / Hno3

Hi TTS, The "Realstorm Engine" is great. Something completely different :o) in the actual scene ;) cu Seppl

ALREADY DONE

ALREADY DONE PLEASE LOOK FOR ANOTHER ONE

Excellent! I've been waiting for someone to actually start to test this level of rendering. Polygons are getting boring....thanks and keep up the great work!

hyllöchen, keep up the good work.. it's still fantastic, and gets bether and bether.. btw, your soft shadows.. how you do them? multisampling? if you don't know this yet, read this: http://www.cs.utah.edu/~bes/papers/ coneShadow/shadow.html or the paper.. http://w3imagis.imag.fr/Membres/Cyril.Soler/ DEA/Ombres/Papers/Parker.pdf anyways, really good work.. can't wait for my p4 3.06 HT, overclocked to i hope quite a bit over 4 giga.. 4.6 or so.. ... uhh that'll be fun to trace rays then:D

This benchmark stuff really kicks ass. Keep up the good work guys. Teh futare si tihs! :)

Great stuff... Nice to see you guys are trying to promote realtime raytracing outside the demoscene. Perhaps it's possible to sideload some of the calculations to the 3D card (bounding boxes, ...)?

Buh! =))))

Wow, very cool. It's nice to see there's actually happening something in the realtime-raytracing world. These techniques will dominate the gaming world in the future.

It doesn't work! I get the error message...cannot set display mode. :-(

I found you...and i'll never leave to make all the future 3D works to my MATROX G400 anymore...With my Athlon XP 2.200 Ghz i'm waiting for a new production of total virtual reality with my favourite engine "realstorm" Karisma

NICK WAS HERE

Hot realtime raytraced action! I can't wait to try this on my new 100 GHz 64-bit Clawhammer. Still awaiting the Gameboy Advance port... Seriously though, cool stuff. This be da future!!

there is no sse,3dnow,mmx yet in? thats all plain c? that is simply amazing! i hope you get in touch with some new compiler like .net everett or the intel compiler to just try it one time (still with the c-code) with sse and 3dnow (sse2 even?:D). would be interested in the possible speedboost.. at least that sounds _very_ promising..

Although I only have a Athlon TB @ 933MHz + 512MB SDR it does not stop me from enjoying your benchmark =) I am eagerly awaiting your next benchmark. Nice work. Respect.

nice work, i see you extremely improved the engine since i got to see a version long time ago :-)

A fast hello to all FAN guys. I never guessed that the engine would end up like it is now when I saw the first version. Hope that the new version will be out soon as the radiosity effects are really promising (even when it is more likely a dia-show than 100 fps) Keep up the good work... see you soon

this this is pouëtish !

Great engine, I was getting tired of polygons, now all you need to do is get this thing into hardware so it's fast enough for games. It's worth to try.

I would like to congratulate everyone who made this app / engine possible. It really shows us how far we are from the graal of 3D performance TOGETHER with quality, even with awesome GPUs from nVidia and ATI. Keep up the good work.

i'm first! i congratulate you, i liked it from the first day i knew of it. i'm active at raytracing myself, want to see this as the future method on gpu's/vpu's.. wish you the best with your project, can't wait for the next demo-benchmark.. (and i _love_ the c-button.. :D) btw, any game in planning? i've seen "the race" but there's nothing about it yet official.. anyways, good night.. i'm first! :D