| Age | Commit message (Collapse) | Author |
|---|
|
The ray tracer was already threaded, so this required little change other
than making some state global like the previous commits, and calling the
underlying non-threaded single-fragment scene renderer function.
A future commit will discard the now vestigial ray_threads related code.
|
|
Same basic changes as the previous commits made to julia and plasma.
|
|
Same general procedure as the previous commit made to the julia module.
|
|
Move maintenance of per-frame variables into julia_prepare_frame(), which
requires making them static file-scope globals for now.
Also make minor adjustments to the code to make less assumptions about the
fragment being rendered (like it's x/y coordinates being 0, etc.)
A future commit will probably add an initializer function to
rototiller_module_t, with an opaque pointer output which will be fed to all
the module methods so these globals can be encapsulated and instantiated.
|
|
This is a simple worker thread implementation derived from the ray_threads
code in the ray module. The ray_threads code should be discarded in a
future commit now that rototiller can render fragments using threads.
If a module supplies a prepare_frame() method, then it is called
per-frame to prepare a rototiller_frame_t which specifies how to divvy
up the page into fragments. Those fragments are then dispatched to a
thread per CPU which call the module's rendering function in parallel.
There is no coupling of the number of fragments in a frame to the number of
threads/CPUs. Some modules may benefit from the locality of tile-based
rendering, so the fragments are simply dispatched across the available CPUs
in a striped fashion.
Helpers will be added later to the fb interface for tiling fragments, which
modules desiring tiled rendering may utilize in their prepare_frame()
methods.
This commit does not modify any modules to become threaded, it only adds
the scaffolding.
|
|
Undecided on wether rototiller_frame_t should be fb_frame_t or not,
may change later.
|
|
Modules need to know the overall dimensions of the frame the fragment
they're rendering is part of. Previously it wasn't really necessary
since the fragments supplied to the modules had always been the full
page, but that's changing.
This commit also changes the julia module to use the frame dimensions,
others will need updating as well.
|
|
Adding more context to the name in anticipation of adding a prepare_frame()
method to the module struct.
|
|
Make consistent with the source directory structure naming.
|
|
|
|
The highlight on the little green sphere was white-washing
the entire thing due to its high specular reflection value.
This produces more reasonable results...
|
|
Was a constant at 20, this allows it to be specified per-object.
|
|
Make spheres a little more diverse in terms of specular/diffusion,
and minor tweak to the plane color.
|
|
|
|
Make the camera orbit around the origin at a varying radius, with kept aimed
facing the origin, with some vertical sweep+tilt thrown in.
|
|
This is unoptimized, with a palette slapped together in vim, but still
pretty neat!
|
|
We should only consider dot products > 0 as intersected, or >= something
very close to 0 (epsilon).
As-is resulted in planes moving with camera movement along the plane normal
axis.
Also fixes plane distance to be non-negative in the current scene.
|
|
Leftover from debugging presumably
|
|
Originally I only implemented pitch->yaw->roll, and being new to all this
didn't fully appreciate the limitation that resulted in.
This adds all six permutations of pitch/yaw/roll, the scene must specify
the desired order when setting up the camera with the euler angles, see
the enum in ray_euler.h.
|
|
trace_ray() bumps the depth, the reflection ray trace_ray() call just needs to
propagate the depth variable not advance it as well. This was probably
vestigial from early development and never got taken out.
This does mean more reflections now, and correspondingly slower rendering, but
it at least makes MAX_RECURSION_DEPTH accurate. The define can be changed if
the performance is too bad.
|
|
|
|
sparkler and stars both cleared fragments and drew individual pixels into fragments, add that functionality to fb.h and cleanup sparkler and stars accordingly.
|
|
|
|
|
|
Currently just a memset wrapper... but maybe could get noop'd if
the flipper thread started supporting pre-zeroing pages in its thread,
just need a zeroed state in the fb page.
|
|
drop draw_pixel() duplication
|
|
|
|
discards draw_pixel(), introduces helpers.h and a convenience
function for bounds checking and oob extermination. Move makergb
to helpers.h, draw.h gets removed in a later commit.
|
|
Modules seem to be duplicating this, just pull it into fb.h since we're
already dependent on the fb.h abstractions in the modules. Slippery slope!
|
|
If a z-buffer is added these checks will need to be done independent from and
prior to drawing. Also it's silly to makergb() pixels which can't be drawn.
|
|
Simple x,y coordinate check
|
|
- Move source to src/ subdir
- Use $(top_srcdir)/src instead of ../../
|
|
The relative path broke out-of-tree builds.
Previously the following:
$ mkdir /tmp/foo
$ cd /tmp/foo
$ ~/src/rototiller/configure
$ make
Would fail to compile unable to locate the headers in
~/rototiller/src
This fixes it.
|
|
Restoring some organizational sanity since adopting autotools.
|
|
If NUM_FB_PAGES was redefined to 1 rototiller would hang, since 1 is
insufficient for page-flipping.
|
|
|
|
|
|
(78d9c385ae3a1939e059c674ba5649df99d5615c)
|
|
As this stuff is all generated we don't want to ever check it in...
|
|
Not everyone knows about `autoreconf --install`
|
|
|
|
Now the usual ./configure && make should suffice.
|
|
Builds were getting too time consuming, autotools is a very simple way
to get incremental builds without having to dick with Makefiles myself.
|
|
stars: add starfield simulator from ph1l/stars
|
|
|
|
vestigial deficiency from variable gravity experiment
also reduced gravity, the results seem more aesthetically pleasing.
|
|
More of 86bc322, eliminate per-particle unnecessary calls
|
|
This is a trivial optimization which makes a substantial difference.
When we're doing things like explosions, thousands of occupants get
added to the bsp tree at the exact same position. Without the lookup
cache we end up traversing the octree down to its maximum depthrepeatedly,
because the bv containing the explosion is of course full and maximally
deep.
Now explosions don't need to spend a bunch of time in the octree just to
keep locating the same full bv the explosion occurs in.
|
|
|
|
trivial per-particle savings
|
