kory@avatar.avatar.com (Kory Hamzeh) (11/27/90)
I wrote a parallel raytracer (named 'prt') a while back with the following features: - Runs on multiple heterogeneous machines networked together using TCP/IP. - Crude load balancing. - Primitives: - Polygons - Spheres - Hallow Sphere - Cones - Cylinders - A object that can be expressed in a quadratic form - Rings - Shading: - Gourard - Phong - Whitted - Rendering: - Simple: one ray per pixel - Stochastic sampling (jitter) - Instances of objects. - Input format: - An extention of nff. I have written a filter which will convert NFF files to prt format. - Output format: - MTV format (24 bit). Note that prt is a parallel raytracer which spawns off children over multiple machines to distribute the work. I have only used it on five Sun Sparcstations and have gotten excellent performance. I'm not aware of any other public domain parallel raytracers other than VM_pRay (which, I believe, runs only on a specific platform). Since there are so many public domain raytracers currently available, I wanted to see if there was any interest in prt before I send it to comp.sources.unix. So, if you are interested in getting your hands on prt, please let me know. If I get enough positive responses, I will post it in c.s.u. Thanks, --kory -- ------------------------------------------------------------------------------- Kory Hamzeh UUCP: avatar!kory or ..!uunet!avatar!kory INTERNET: kory@avatar.com
kory@avatar.avatar.com (Kory Hamzeh) (11/28/90)
In article <105@avatar.avatar.com>, kory@avatar.avatar.com (Kory Hamzeh) writes: > > > I wrote a parallel raytracer (named 'prt') a while back with the following > features: > [ lots of stuff deleted ] > > Since there are so many public domain raytracers currently available, > I wanted to see if there was any interest in prt before I send it > to comp.sources.unix. > > So, if you are interested in getting your hands on prt, please let me > know. If I get enough positive responses, I will post it in c.s.u. > > ------------------------------------------------------------------------------- > Kory Hamzeh UUCP: avatar!kory or ..!uunet!avatar!kory > INTERNET: kory@avatar.com I received over 100 request for prt within 24 hours of posting the above article. I can't mail to everyone for obvious reasons and if I post it to comp.sources.unix, it might take up to 1 month before Rich Salze can release it (Rich has a very busy schedule). I can do any of the following: - Make it available via anonymous uucp - Post it to comp.graphics - I don't have ftp access, but if some who does can make it available for anonymous ftp, I can mail them a copy. (If you are interested in archiving prt for anonymous ftp access, please let me know). Posting to comp.graphics would be the fastest way, but I have a funny feeling it will piss off some of the net gurus. Either way, I need to beef up the documentation before I release it. I hope to send it out by the end of the week. And lastly, I will be upgrading the OS on my mail gateway over the weekend and I expect to have some minor e-mail glitches. If you send me mail and it bounces, please try again. Anyway, let me know. Thanks, --kory -- ------------------------------------------------------------------------------- Kory Hamzeh UUCP: avatar!kory or ..!uunet!avatar!kory INTERNET: kory@avatar.com
marcel@tnoibbc.UUCP (Marcel Boender) (11/28/90)
Article 11961 Kory Hamzeh writes : > ... > So, if you are interested in getting your hands on prt, please let me > know. If I get enough positive responses, I will post it in c.s.u. Yes, please post it ! -- Marcel Boender : TNO-BOUW INTERNET : marcel@tnoibbc : PO-box 49 DOMAIN : marcel@ibbc.tno.nl : 2600 AA Delft UUCP : ..!hp4nl!tnoibbc!marcel : the Netherlands PHONE : +31 15 842031
garyc@hobbes.wv.tek.com (Gary Combs) (11/29/90)
In article <2188@tnoibbc.UUCP>, marcel@tnoibbc.UUCP (Marcel Boender) writes: |> From: marcel@tnoibbc.UUCP (Marcel Boender) |> Subject: Re: Any interest in a parallel raytracer? |> Keywords: Parallel Raytracers |> Date: 28 Nov 90 13:31:14 GMT |> |> Article 11961 Kory Hamzeh writes : |> |> > ... |> > So, if you are interested in getting your hands on prt, please let me |> > know. If I get enough positive responses, I will post it in c.s.u. |> |> Yes, please post it ! |> -- |> Marcel Boender : TNO-BOUW INTERNET : marcel@tnoibbc |> : PO-box 49 DOMAIN : marcel@ibbc.tno.nl |> : 2600 AA Delft UUCP : ..!hp4nl!tnoibbc!marcel |> : the Netherlands PHONE : +31 15 842031 -- Dito for me. ------------------------------------------------------------------------------- Gary Combs Tektronix, Inc NWD Technical Marketing Support (503)685-2072 garyc@shamu.wv.tek.com -------------------------------------------------------------------------------
jonas-y@isy.liu.se (Jonas Yngvesson) (11/30/90)
kory@avatar.avatar.com (Kory Hamzeh) writes: >I wrote a parallel raytracer (named 'prt') a while back with the following >features: I have made a network distributed version of David K. Buck's ray-tracer DKB. > - Runs on multiple heterogeneous machines networked together > using TCP/IP. Same here, I used inetd to fire up a trace server on request from the client. The client tokenizes the input file and sends these tokens to each server. The client then distributes the job, one scanline per server until the image is finished. > - Crude load balancing. > - Primitives: > - Polygons > - Spheres > - Hallow Sphere > - Cones > - Cylinders > - A object that can be expressed in a quadratic form > - Rings > - Shading: > - Gourard > - Phong > - Whitted > - Rendering: > - Simple: one ray per pixel > - Stochastic sampling (jitter) > - Instances of objects. > - Input format: > - An extention of nff. I have written a filter which > will convert NFF files to prt format. DDKB (distributed dkb) has the same primitives, shading and input format as DKB (for obvious reasons). This includes very nice support for solid texturing. Antialiasing differs slightly. DKB uses an adaptive jittered supersampling, but in DDKB, each server only has access to one scanline at a time. This means adaption is done in x-direction only. > - Output format: > - MTV format (24 bit). Here I have used a mix between the MTV-format an the QRT-format used in DKB. An MTV-header is followed by the scanlines in QRT-format. Each line tagged with its line number (this is because the lines are stored in the order they are finished by the servers and must be sorted before display). >Note that prt is a parallel raytracer which spawns off children over >multiple machines to distribute the work. I have only used it on five >Sun Sparcstations and have gotten excellent performance. I'm not aware >of any other public domain parallel raytracers other than VM_pRay >(which, I believe, runs only on a specific platform). Yeah! We have tried DDKB running on 55 sparcstations (then we ran out of file descriptors, perhaps we should use UDP instead of TCP). Pretty good performance indeed. Unfortunately DKB is not a terribly fast tracer in itself and I have no time to hack around in it. I'm not very willing so send this thing out though. Partly because it is still only a hack, and partly because I have not contacted David Buck and asked what he thinks about the whole thing. --Jonas ------------------------------------------------------------------------------ J o n a s Y n g v e s s o n Dept. of Electrical Engineering jonas-y@isy.liu.se University of Linkoping, Sweden ...!uunet!isy.liu.se!jonas-y -- ------------------------------------------------------------------------------ J o n a s Y n g v e s s o n Dept. of Electrical Engineering jonas-y@isy.liu.se University of Linkoping, Sweden ...!uunet!isy.liu.se!jonas-y
valdis@wizards.vt.edu (Valdis Kletnieks) (11/30/90)
In article <105@avatar.avatar.com>, kory@avatar.avatar.com (Kory Hamzeh) writes: |> Since there are so many public domain raytracers currently available, |> I wanted to see if there was any interest in prt before I send it |> to comp.sources.unix. |> |> So, if you are interested in getting your hands on prt, please let me |> know. If I get enough positive responses, I will post it in c.s.u. Kory: (1) I'm interested, as I'm sure a lot of other people are as well. However, please note that (2) comp.sources.unix is a "black hole" - the moderator seems to have fallen asleep at the wheel, and nothing posted to it ever surfaces. You may wish to consider posting it to this newsgroup instead? Valdis Kletnieks
kjartan@puce.inria.fr (Kjartan Emilson) (11/30/90)
Does anybody have an algorithm for a massively parallel raytracer, which could for example run on a Connection Machine, i.e a machine with a 'single instruction - many data' architecture ? -Kjartan
kyriazis@iear.arts.rpi.edu (George Kyriazis) (12/01/90)
In article <9308@mirsa.inria.fr> kjartan@puce.inria.fr (Kjartan Emilson) writes: > >Does anybody have an algorithm for a massively parallel raytracer, >which could for example run on a Connection Machine, i.e a machine >with a 'single instruction - many data' architecture ? > The ray-tracing algorithm is usually paralellelized ray-at-a-time, which makes it suitable for MIMD machines. If you want to run it on an SIMD you have a problem, since you have to advance all rays at the same time, then filter out which rays do not need a second reflection, and trace the second level, etc. I believe that Scott Whitman (at uiuc?) has written a ray-tracer for a SIMD machine. I don't remember any details. Pick up the SIGGraph proceedings for either '88 or '89 (don't remember which one) on ray-tracing or parallel computer architectures for computer graphics and you are going to see his name there. Hope that helps. ---------------------------------------------------------------------- George Kyriazis kyriazis@rdrc.rpi.edu kyriazis@iear.arts.rpi.edu
punch@pleiades.cps.msu.edu (Bill Punch) (12/04/90)
I know of one researcher who did work on implementation of ray-tracing
on a massively parallel machine like the Connection Machine. I got it as
a tech report from the connection machine peopleT(TMC), it is also a
tech report at the MIT Media lab.
Hubert Delany, "Ray Tracing on a Connection Machine", Tech report from
Thinking Machines Corp., Tech Report Number VZ88-3, April 1988.
Very Intersting Stuff!
>>>bill punch<<<
punch@pleiades.cps.msu.edu
mike@BRL.MIL (Michael John Muuss <mike>) (12/04/90)
In article <105@avatar.avatar.com>, kory@avatar.avatar.com (Kory Hamzeh) writes: |> I wrote a parallel raytracer (named 'prt') a while back ... |> - Runs on multiple heterogeneous machines networked together |> using TCP/IP. |> - Crude load balancing. Well, the BRL-CAD Package ray-tracer operates both on shared memory multi-processors, as well as in a network-distributed mode. While not PD, it is free. I have written about it in a number of articles, including: %K MUUS90a %A M. J. Muuss %T Workstations, Networking, Distributed Graphics, and Parallel Processing %E D. F. Rogers, R. A. Earnshaw %B Computer Graphics Techniques: Theory and Practice %I Springer-Verlag %D 1990 While not "super sophisticated", the load balancing algorithm is non-trivial. It assigns pixel blocks of variable size. Three assignments are outstanding to each worker at any time, to pipeline against network delay. New assignment size is tuned, based upon measured past performance (weighted average with variable weighting, depending on circumstances). I'd certainly be interested in seeing a copy of your work. Best, -Mike Muuss Advanced Computer Systems Ballistic Research Laboratory APG, MD 21005-5066
kory@avatar.avatar.com (Kory Hamzeh) (12/06/90)
In article <14648@smoke.brl.mil>, mike@BRL.MIL (Michael John Muuss <mike>) writes: > > [ lots of good stuff about BRL-CAD deleted ] > > While not "super sophisticated", the load balancing algorithm is > non-trivial. It assigns pixel blocks of variable size. Three > assignments are outstanding to each worker at any time, to > pipeline against network delay. New assignment size is tuned, > based upon measured past performance (weighted average with variable > weighting, depending on circumstances). > > I'd certainly be interested in seeing a copy of your work. > Prt's load balancing is not as sophisticated as BRL-CAD's. Prt simply multiplexes the scanlines across the different machines. The slowest machine on the network will be the bottle neck. When I get a chance, I would like to use the same techniques used in BRL-CAD. I think that it is the best load balancing for a raytracer. --kory -- ------------------------------------------------------------------------------- Kory Hamzeh UUCP: avatar!kory or ..!uunet!avatar!kory INTERNET: kory@avatar.com