Chunky by FelipeFS
Chunky by FelipeFS
GPWiki.org
It is currently Fri Sep 19, 2014 3:49 pm

All times are UTC




Post new topic Reply to topic  [ 19 posts ] 
Author Message
PostPosted: Fri Jan 11, 2013 3:36 am 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
I bought an AMD FX-8150 last year particulary in excitement to use it for future software rasterization experiments. I haven't been able to try any of my software rendering ideas yet, but I'm probably going to start soon (after I finish my current projects). I'm thinking that this processor will be ideal for software rasterization, would you agree? Based on the described bulldozer architecture, it's supposed to be very good for integer math and most people recommend using fixed-point arithmetic for rasterization. That's really the reason I chose to buy it. What do you guys think about this decision?

Edit:
Nothing to say? Hmm. *bored* :x

Anybody else have any interesting plans? I'd like to discuss your ideas. :)


Top
 Profile  
 
PostPosted: Sun Feb 03, 2013 5:21 pm 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
Why do you want to write a software renderer though? just for experiments sake?

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Sun Feb 03, 2013 8:00 pm 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
No, I want to make some fundamental changes to what primitives are that likely cannot be achieved through conventional graphics programming (i.e. even modern shader models). There are still numerous advantages to working with the CPU, at least in the first place. If I have any success then I might transfer my work to a GPGPU basis. :)


Top
 Profile  
 
PostPosted: Tue Feb 05, 2013 3:04 am 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
what kind of fundamental changes :P

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Thu Feb 07, 2013 5:13 pm 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
I have been very excited about these indicible ideas for more than 18 months, but I just recently decided to not share them. A small number of researchers have realized very insufficient glimpses of this theoretical ensemble. I don't trust their understanding to appropriately comprehend the theoretical magnitude of this project.

Regardless of how much conceit you suspect I demonstrate here, I will state my perception that many well-respected graphics researchers have unsatisfactorily exhibited naive comprehension and senseless heuristics. I've decided I do not want to share these ideas until I completely implement them properly and to their fullest; myself now understanding how such magnitudes of insight are usually misinterpreted and abused during senseless adoption practices. When I release it, I will try to introduce the foundational concepts and insights for an effectively lucid reception.

This, however, is not a project of application. It is theoretical -- inherently it involves a lot more than merely fashioning a new paradigm of primitive representation -- so experimenting in the domain of software rasterization is only one part of my project.

For this topic, I intended to discuss CPU rasterization in general, both from algorithmic standpoints and a discussion analyzing hardware characteristics/implications with regards to software rendering.


Top
 Profile  
 
PostPosted: Thu Feb 07, 2013 7:01 pm 
Grand Optimizer

Joined: Sun Oct 16, 2011 3:09 pm
Posts: 364
Location: Here (where else?)
I always wonder why people post about having a secret :)

_________________
My project: Messing about in FreeRCT, dev blog, and IRC #freerct at oftc.net


Top
 Profile  
 
PostPosted: Thu Feb 07, 2013 8:53 pm 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
He asked me. I never expected to give an answer; it's off-topic.

Quote:
For this topic, I intended to discuss CPU rasterization in general, both from algorithmic standpoints and a discussion analyzing hardware characteristics/implications with regards to software rendering.


Top
 Profile  
 
PostPosted: Thu Feb 07, 2013 9:29 pm 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
And yet you tell us about it! You expect us to not be curious?

I don't know your background and skillset, but judging from your posts i can see that you've got a good perspective about graphics programming generally. Still making claims like that makes you appear like a quack.
Not that i want to diminish your rep or knowledge, i just don't know you any better to know what i can expect from a claim like that.

Reminds me of the Unlimited Detail claim that's been going around for a few years, which appears to be a glorified voxel/point cloud renderer.

On Topic though: I haven't worked with rasterization in detail. I know how it basically works technically, but my knowledge about it is just about good enough to be able to push vertices around on the gpu.

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Thu Feb 07, 2013 11:12 pm 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
Quote:
Still making claims like that makes you appear like a quack.

I didn't want to go this direction in the first place, but I also didn't want to leave you hanging. What else could I have done? Also, I did not make any claims in the first place. I merely explained why I didn't feel comfortable talking about it. It's quite deep and my most recent experiences with sharing ideas have taught me to fear arrogance.


Top
 Profile  
 
PostPosted: Fri Feb 08, 2013 12:12 am 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
Fair enough.
Although you just could have said it's for experiments sake (which it basically is anyways) and i'd have been satisfied by that ;)

If you don't wanna share it at this point thats fine. Do whatever you feel you have to do. I'll be looking forward to the details and results.

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Fri Feb 08, 2013 3:22 am 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
Cool. I'm just going to spam related information to draw it back on topic. :D

http://software.intel.com/en-us/article ... -larrabee/
http://fgiesen.wordpress.com/2013/02/06 ... conspirac/
http://devmaster.net/forums/topic/1145- ... erization/
http://www.gamedev.net/blog/1390/entry- ... r-updates/ http://cse.taylor.edu/%7Ezbethel/MSR/Mo ... chToSR.pdf

Interesting... eh? :O


Top
 Profile  
 
PostPosted: Sat Feb 09, 2013 6:53 am 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
So, I've been looking for tools to help develop this project as fast as possible.

Here's pixel toaster, a library designed for efficiently displaying the output of software renderers:
http://code.google.com/p/pixeltoaster/

Does anybody know of high quality & free-software math libraries designed for parallel computing, or anything else useful for that matter?


Top
 Profile  
 
PostPosted: Sat Feb 09, 2013 3:29 pm 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
Are you going to use the opencl/cuda for your calculations?
You could also write your renderer in software and still render the buffer through opengl or directx.

The only math library i've ever used is slim-math (c#) but this one looks fairly feature complete. Although i'm not sure how well it is optimized.

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Sat Feb 09, 2013 4:44 pm 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
*snip* Nevermind. Yes I am writing it in software. No, I am not using C#...
I decided to not use pixel toaster. I really don't need it...


Top
 Profile  
 
PostPosted: Tue Feb 12, 2013 3:25 am 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
i guessed as much that you wouldn't use c#, but it is easy to port to c/cpp or whatever else. but i'd assume there are enough decent c/cpp libs out there aswell

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Tue Apr 23, 2013 5:59 am 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
OpenCL FTW! :D


Top
 Profile  
 
PostPosted: Tue Apr 23, 2013 2:04 pm 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
really! :D

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
PostPosted: Tue Apr 23, 2013 10:49 pm 
BANNED

Joined: Sun Jun 24, 2012 12:49 am
Posts: 504
Yehp.

Image


Top
 Profile  
 
PostPosted: Tue Apr 23, 2013 11:39 pm 
Digerati

Joined: Thu Sep 09, 2004 1:17 pm
Posts: 1814
Location: burrowed
Fancy!

_________________
Long pork is people!

wzl's burrow


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 19 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  

Powered by phpBB® Forum Software © phpBB Group