Simple profiler scopes for wgpu using timer queries
tracy feature flag)TODO: * Better error messages * Disable via feature flag
Create a new profiler object:
use wgpu_profiler::{wgpu_profiler, GpuProfiler};
// ...
let mut profiler = GpuProfiler::new(4, queue.get_timestamp_period(), device.features()); // buffer up to 4 frames
Using scopes is easiest with the macro:
wgpu_profiler!("name of your scope", &mut profiler, &mut encoder, &device, {
// wgpu commands go here
});
Note that GpuProfiler reads the device features - if your wgpu device doesn't have wgpu::Features::TIMESTAMP_QUERY enabled, it will automatically not attempt to emit any timer queries.
Similarly, if wgpu::Features::WRITE_TIMESTAMP_INSIDE_PASSES is not present, no queries will be issued from inside passes.
Wgpu-profiler needs to insert buffer copy commands, so when you're done with an encoder and won't do any more profiling scopes on it, you need to resolve the queries:
profiler.resolve_queries(&mut encoder);
And finally, to end a profiling frame, call end_frame. This does a few checks and will let you know if something is off!
profiler.end_frame().unwrap();
Retrieving the oldest available frame and writing it out to a chrome trace file.
if let Some(profiling_data) = profiler.process_finished_frame() {
// You usually want to write to disk only under some condition, e.g. press of a key or button
wgpu_profiler::chrometrace::write_chrometrace(std::path::Path::new("mytrace.json"), &profiling_data);
}
To get a look of it in action, check out the example project!
Licensed under either of
at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.
$ claude mcp add wgpu-profiler \
-- python -m otcore.mcp_server <graph>