Nvidia Demonstrates Interactive Ray-tracing

Status
Not open for further replies.

dragonsqrrl

Distinguished
Nov 19, 2009
1,280
0
19,290
3
Rendering scenes using ray-tracing and global illumination takes freakin FOREVER, even on the i7 based workstations at school. I'm amazed at how quickly you're able to preview a scene using the hardware acceleration provided by Nvidia's latest generation of GPU's. It's literally just a matter of seconds based on the videos I've seen.

This really does make "interactive" ray-tracing possible for the first time on a desktop... awesome. This is the sort of application the Fermi architecture really excels at. The emphasis Nvidia places on this market is probably the main reason I'll be going with a Fermi based solution for my next build, and not an Evergreen.
 

Blessedman

Distinguished
May 29, 2001
577
0
18,980
0
I remember back in the day (late 80's) when Amiga had it's big hay day with ray tracing, there were pundits that said that real time ray tracing would never be a reality. This was back when a single frame would take days to render in a farm. Oh how far we have come.
 
i gave Maya a try before, fun, but needs time to learn. and yes it takes long to render a realistic scene, add that to my not-so-good settings being a noob and all :)
this is good news. so are there 3d modeling applications that makes use of the gpu? maya/3ds?
 

Draven35

Distinguished
Nov 7, 2008
806
0
19,010
9
The F-18 frame is one frame while someone was interacting with it. The image refines as it rests. The Bunkspeed shot of the viper probably took a minute, maybe two. Before iray 'refines' the image, you can still get an idea for how the lighting and GI will look, well enough to make lighting decisions.
 

matt87_50

Distinguished
Mar 23, 2009
1,150
0
19,280
0
[citation][nom]dragonsqrrl[/nom]Rendering scenes using ray-tracing and global illumination takes freakin FOREVER, even on the i7 based workstations at school. I'm amazed at how quickly you're able to preview a scene using the hardware acceleration provided by Nvidia's latest generation of GPU's. It's literally just a matter of seconds based on the videos I've seen. This really does make "interactive" ray-tracing possible for the first time on a desktop... awesome. This is the sort of application the Fermi architecture really excels at. The emphasis Nvidia places on this market is probably the main reason I'll be going with a Fermi based solution for my next build, and not an Evergreen.[/citation]

i7 970 extreme (the old quad core one) = 48 Gflops, these video cards = 1000 to 2000 Gflops.

pwned.



any demos of real time raytracing? (20, 30fps, reasonable res ect)

obviously images like the above which look REAL aren't gonna be real time yet, but some form of awesome raytracing might be.
 

matt87_50

Distinguished
Mar 23, 2009
1,150
0
19,280
0
hey, that viper pic: is it just the car that is rendered, then composited with a picture? or is the whole terrain rendered too? I could scarcely believe that...
 

lukeeu

Distinguished
Mar 12, 2010
75
0
18,630
0
[citation][nom]matt87_50[/nom]i7 970 extreme (the old quad core one) = 48 Gflops, these video cards = 1000 to 2000 Gflops.pwned.any demos of real time raytracing? (20, 30fps, reasonable res ect)obviously images like the above which look REAL aren't gonna be real time yet, but some form of awesome raytracing might be.[/citation]
The problem with ray tracing isn't computational power. It's memory access. Every reflection takes only ~10 arithmetic operations to compute and then you need to find next intersection ant this is where problem starts.For a 1m triangle scene you'll have to do at least 30 reads from your data structure or even 1000s in worse case scenarios. Every read from RAM on a graphics card takes hundreds of clock cycles also the reads are random and not sequential so it won't hit very small caches and you get x16 RAM slowdown for misalignment. On a CPU you get megabytes of cache and no memory access restrictions. These algorithms work in O(n) and O(n lg n) times so you need to feed about as much data to the GPU as it can compute so for a 1Tflop chip you should connect memory that can do 1Tflop * 4 bytes = 4TB/s memory @ random reads... this can be only done with on-chip cache and for a 1m triangle scene you will need ~ 1000000triangles * 3 vertexes * 12bytes/vertex + 1 to 8 million nodes of data structure * (8 children + 1 parent) *4 bytes = way over 60MB of cache and this probably will have to be accessed by hundreds of PUs simultaneously... so it should be kept in dozens of copies.
 

matt87_50

Distinguished
Mar 23, 2009
1,150
0
19,280
0
[citation][nom]lukeeu[/nom]The problem with ray tracing isn't computational power. It's memory access. Every reflection takes only ~10 arithmetic operations to compute and then you need to find next intersection ant this is where problem starts.For a 1m triangle scene you'll have to do at least 30 reads from your data structure or even 1000s in worse case scenarios. Every read from RAM on a graphics card takes hundreds of clock cycles also the reads are random and not sequential so it won't hit very small caches and you get x16 RAM slowdown for misalignment. On a CPU you get megabytes of cache and no memory access restrictions. These algorithms work in O(n) and O(n lg n) times so you need to feed about as much data to the GPU as it can compute so for a 1Tflop chip you should connect memory that can do 1Tflop * 4 bytes = 4TB/s memory @ random reads... this can be only done with on-chip cache and for a 1m triangle scene you will need ~ 1000000triangles * 3 vertexes * 12bytes/vertex + 1 to 8 million nodes of data structure * (8 children + 1 parent) *4 bytes = way over 60MB of cache and this probably will have to be accessed by hundreds of PUs simultaneously... so it should be kept in dozens of copies.[/citation]

very true

thats why triangles are lame. they are for lame rasterizers. all shapes in raytracing should be formed from much more complex geometric objects :D

for instance: terrain: should just a the equation for the plasma fractal you would otherwise use to generate a hight map. and bam! like 10s of bytes for your whole terrain! pwned!

I mean, the ray collision equation might be a *bit* more complex.... but Tflops!!

like how they are trying to overcome the same memory limitations in realtime rendering with tessellation.
 

lukeeu

Distinguished
Mar 12, 2010
75
0
18,630
0
[citation][nom]weirdguy99[/nom]That image of the Viper is bloody amazing.[/citation] Viper looks HORRIBLE for ray tracing!
-No car shadow reflected in body
-No road and shadow reflections on the rims
-No reflection of the mirror in the windows
-No reflection of the scenery in the glass
-No double or triple reflections
Only reflections that wouldn't be done in a super easy traditional way:
-Mirror in the bodywork (but could be done with some effort)
-Scenery reflection in the lamp shadowed by the lamp (could be done)
-Shadow on the brakes (could use a dedicated shadow here)

Only thing that looks great here is the scenery but it's only because of the detail not ray tracing. It could be rendered using Oblivion level shaders..
 

gpace

Distinguished
Mar 5, 2010
69
0
18,630
0
I can't wait till technology allows this to be real time. I'd also like a side of realistically destructible environment and a engaging story.

Good job Nvidia and I hope ATI is working on something like this too. Nothing like some good competition to push technology forward.
 

fonzy

Distinguished
Dec 23, 2005
398
1
18,785
0
[citation][nom]gpace[/nom]I can't wait till technology allows this to be real time. I'd also like a side of realistically destructible environment and a engaging story.Good job Nvidia and I hope ATI is working on something like this too. Nothing like some good competition to push technology forward.[/citation]

I hope so to,ATI's next set of cards are do out later this year hopefully they have something that can beat Nvidia.
 

jednx01

Distinguished
Mar 31, 2008
448
0
18,810
7
Wow... I have to admit that this is one of the most impressive jobs of computer generated images that I've ever seen. I seriously can't tell that the picture of the viper isn't a real photo... :O
 

demonhorde665

Distinguished
Jul 13, 2008
1,492
0
19,280
0
[citation][nom]Cons29[/nom]i gave Maya a try before, fun, but needs time to learn. and yes it takes long to render a realistic scene, add that to my not-so-good settings being a noob and all this is good news. so are there 3d modeling applications that makes use of the gpu? maya/3ds?[/citation]

as far as i know of it ,no 3d apps make use of teh gpu past rendering teh basic view port , this is because when go to "render" a scene it appliies all the mapping technologies with teh cpu becuase a lot of those technologies can not be done on a video card (which uses rasterization) namely ray tracing , photometric lights, and ray traced shadows, by the time any of tehse tecnologies are doable by a video card ,they will likely have newer technologies that can't be rendered by a video card
 

cinergy

Distinguished
May 22, 2009
251
0
18,780
0
[citation][nom]weirdguy99[/nom]That image of the Viper is bloody amazing.[/citation]

I can use 3d-studio too. Rendering solid objects like that is no big deal.
 
Status
Not open for further replies.

ASK THE COMMUNITY