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/til.c | |
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/til.c')
-rw-r--r-- | src/til.c | 16 |
1 files changed, 8 insertions, 8 deletions
@@ -108,9 +108,9 @@ void til_shutdown(void) } -static void _blank_prepare_frame(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment, til_fragmenter_t *res_fragmenter) +static void _blank_prepare_frame(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment, til_frame_plan_t *res_frame_plan) { - *res_fragmenter = til_fragmenter_slice_per_cpu; + *res_frame_plan = (til_frame_plan_t){ .fragmenter = til_fragmenter_slice_per_cpu }; } @@ -182,22 +182,22 @@ static void module_render_fragment(til_module_context_t *context, til_threads_t module = context->module; if (context->n_cpus > 1 && module->prepare_frame) { - til_fragmenter_t fragmenter; + til_frame_plan_t frame_plan = {}; - module->prepare_frame(context, ticks, fragment, &fragmenter); + module->prepare_frame(context, ticks, fragment, &frame_plan); if (module->render_fragment) { - til_threads_frame_submit(threads, fragment, fragmenter, module->render_fragment, context, ticks); + til_threads_frame_submit(threads, fragment, &frame_plan, module->render_fragment, context, ticks); til_threads_wait_idle(threads); } } else if (module->prepare_frame) { - til_fragmenter_t fragmenter; + til_frame_plan_t frame_plan = {}; unsigned fragnum = 0; til_fb_fragment_t frag; - module->prepare_frame(context, ticks, fragment, &fragmenter); + module->prepare_frame(context, ticks, fragment, &frame_plan); - while (fragmenter(context, fragment, fragnum++, &frag)) + while (frame_plan.fragmenter(context, fragment, fragnum++, &frag)) module->render_fragment(context, ticks, 0, &frag); } else if (module->render_fragment) module->render_fragment(context, ticks, 0, fragment); |