Archives 2009
Roughing in the Play in Design section
Written by
on
in
Design Theory.
My point-form outline for my 1997 thesis was 30 pages of single-spaced 10pt text. It's somewhat daunting going through it and picking out which pieces need to be in a document for first-year students. For tonight I dropped in the "Play in Design" section as a fairly quick rough draft. I figure it eventually deserves ...
Sphere rendering with numpy...
Written by
on
in
Snaking.
So, you ask; "how do you generate the coordinates necessary to render a (unit) sphere" (you do ask the strangest questions)? Well, since you insist, here's a quick rundown of one way to go about it. We'll assume that we're going to use a pair of vertex/index VBOs to render the geometry. We won't try ...
Do not use "profile", use "cProfile"
Written by
on
in
Snaking.
For those who didn't see the presentation at PyCon:
- do not use the profile module with python 2.x
- use the cProfile module (or hotshot if you're on < 2.4)
the profile module has extremely high overhead and will produce results that would have you doing the wrong thing if you were to believe them. See ...
Most of the work I planned to do is done...
Written by
on
in
Snaking.
There's always going to be lots more work to get done, but I think I'm pretty much okay with where OpenGL_accelerate is for now. There's now a C-level API for format handlers, so that Cython modules can be written that pretty much do a cast and a dereference to get values, rather than doing lots ...
Where are we...
Written by
on
in
Snaking.
So with a fairly extensive set of Cython wrappers, running a fairly large, heavy VRML world with transparencies and all the rest, I see a noticeable, but modest, speedup.
The world I'm testing on goes from ~12.8fps (avg) to 15.5fps (avg), that's a nice 20% speedup. On average OpenGLContext now spends around 20% of rendering ...
Cython for the last couple of days...
Written by
on
in
Snaking.
I've been writing accelerator modules for PyOpenGL for the last two days. I think I've got to a reasonably comfortable point. I've got a Numpy FormatHandler and an ArrayDatatype written. With some judicious use of cdef'd typed self variables there's a lot of overhead eliminated from the wrapper operation. To make it really fast, however, ...
Another round of fixing suspend
Written by
on
in
Tuxedo.
Seems that once again I've lost suspend with an Ubuntu update. Was working perfectly well with Intrepid, but Jaunty just gives me the old, familiar black screen on resume. Some day this silly laptop will resume properly out-of-box, but today is not that day.
First step in eliminating GLU quadrics...
Written by
on
in
Snaking.
OpenGLContext is currently using GLU quadrics to render the various VMRL quadric primitives. There's no particular reason to do so other than it was easy and simple to implement. I've just written up some code to generate two VBO-compatible arrays that render a sphere with a given angular resolution as a single glDrawElements call (once ...
GL_ARB_compatibility
Written by
on
in
Snaking.
So yes, apparently the ARB listens when its remaining community screams. This extension (as Lorenzo points out) has been promised to be supported by nVidia on all of its planned 3.x cards and it effectively "undoes" the deprecation of the legacy APIs in OpenGL 3.1. So you can likely get away with not rewriting, as ...
Profiler modules for fun and profit...
Written by
on
in
Snaking.
I wrote a silly little profiler module yesterday (that is, an actual profiler, not a profile viewer). It's currently a Python-coded profiler, so pretty much useless, but it was very fun to write :) . It captures the same basic data as cProfile (that is, it does run-time compression of the data-set rather than logging ...
Monthly archives
- January 2009
- February 2009
- March 2009
- April 2009
- May 2009
- June 2009
- July 2009
- August 2009
- September 2009
- October 2009
- November 2009
- December 2009