StageProfiler.hpp File Reference

Env-gated per-stage wall-time accumulator for the GPU-direct video pipelines. More...

Detailed Description

Env-gated per-stage wall-time accumulator for the GPU-direct video pipelines.

Zero overhead unless SCORE_AJA_PROFILE=1 is set in the environment (one branch on a cached bool). Each profiled stage prints its running mean every printEvery samples to stderr:

[prof] texgen-paint                 n=240 mean=41.273 ms

Intended for the roundtrip harnesses and one-off perf investigations — not a tracing system. For GPU-side truth use nsys; this attributes render-thread wall time to pipeline stages, which is what gates fps when the producer is the bottleneck.

Go to the source code of this file.

Classes
struct	score::gfx::interop::StageProfiler
struct	score::gfx::interop::StageProfiler::Scope

Namespaces
namespace	score
	Base toolkit upon which the software is built.
namespace	score::gfx
	Graphics rendering pipeline for ossia score.

Macros
#define	SCORE_STAGE_PROFILE(var, label)

Macro Definition Documentation

SCORE_STAGE_PROFILE

#define SCORE_STAGE_PROFILE	(		var,
			label
	)

Value:

  static ::score::gfx::interop::StageProfiler var{label};     \
  ::score::gfx::interop::StageProfiler::Scope var##_scope { var }

Profile the enclosing scope under label. One static accumulator per call site; prints running means every 240 samples when enabled.