diff options
| author | Vito Caputo <vcaputo@pengaru.com> | 2022-06-10 20:11:23 -0700 |
|---|---|---|
| committer | Vito Caputo <vcaputo@pengaru.com> | 2022-06-10 21:22:09 -0700 |
| commit | d5db150801023c307fcbe1cd98b4fd8e2d27b55e (patch) | |
| tree | 13657dafe98f8e707760fe176a9fa0bd05866b1f /src/modules/roto | |
| parent | 852ccfb6ffef113003378526c422e591d9339b85 (diff) | |
til: introduce til_frame_plan_t and .cpu_affinity
modules/checkers w/fill_module=$module requires a consistent
mapping of cpu to fragnum since it creates a per-cpu
til_module_context_t for the fill_module.
The existing implementation for threaded rendering maximizes
performance by letting *any* scheduled to run thread advance
fragnum atomically and render the acquired fragnum
indiscriminately. A side effect of this is any given frame, even
rendered by the same module, will have a random mapping of
cpus/threads to fragnums.
With this change, the simple til_module_t.prepare_frame() API of
returning a bare fragmenter function is changed to instead return
a "frame plan" in til_frame_plan_t. Right now til_frame_plan_t
just contains the same fragmenter as before, but also has a
.cpu_affinity member for setting if the frame requires a stable
relationship of cpu/thread to fragnum.
Setting .cpu_affinity should be avoided if unnecessary, and that
is the default if you don't mention .cpu_affinity at all when
initializing the plan in the ergonomic manner w/designated
initializers. This is because the way .cpu_affinity is
implemented will leave threads spinning while they poll for
*their* next fragnum using atomic intrinsics. There's probably
some room for improvement here, but this is good enough for now
to get things working and correct.
Diffstat (limited to 'src/modules/roto')
| -rw-r--r-- | src/modules/roto/roto.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/src/modules/roto/roto.c b/src/modules/roto/roto.c index 1831d2b..a8ee45d 100644 --- a/src/modules/roto/roto.c +++ b/src/modules/roto/roto.c @@ -174,7 +174,7 @@ static void init_roto(uint8_t texture[256][256], int32_t *costab, int32_t *sinta /* prepare a frame for concurrent rendering */ -static void roto_prepare_frame(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment, til_fragmenter_t *res_fragmenter) +static void roto_prepare_frame(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment, til_frame_plan_t *res_frame_plan) { roto_context_t *ctxt = (roto_context_t *)context; static int initialized; @@ -185,7 +185,7 @@ static void roto_prepare_frame(til_module_context_t *context, unsigned ticks, ti init_roto(texture, costab, sintab); } - *res_fragmenter = til_fragmenter_slice_per_cpu; + *res_frame_plan = (til_frame_plan_t){ .fragmenter = til_fragmenter_slice_per_cpu }; // This governs the rotation and color cycle. if (ticks != context->ticks) { |