tree: ed2b9a2e06547ad0aa8124cd22428c81d4d55f97 [path history] [tgz]
  1. BUILD.gn
  2. README.md
  3. flutter_processing_helpers.go
  4. gfx_benchmarks_standalone.sh
  5. process_gfx_trace.go
  6. process_trace_helpers.go
  7. run_scenic_benchmark.sh
  8. scenic_benchmark_example.sh
  9. scenic_processing_helpers.go
  10. trace_from_host.sh
  11. vulkan_is_supported/
garnet/bin/ui/tests/performance/README.md

Scenic Benchmarks

We collect benchmarks first by tracing a given process, and then by post-processing the trace to output benchmark information.

There is only one benchmark here currently but more will be added.

hello_scenic_benchmark.sh

This script collects a trace of hello_scenic. It takes one parameter, which is where the outputted benchmarks JSON file should be written.

process_scenic_trace.go

A Go program that takes the following parameters in this order:

  • test label
  • trace file name
  • output filename for benchmarks The output is a JSON file with the benchmark output.

Example Usage

fx shell "/pkgfs/packages/scenic_benchmarks/0/bin/gfx_benchmarks_standalone.sh /tmp \
  --catapult-converter-args \
  --bots fake_bot_name \
  --masters fuchsia.try \
  --execution-timestamp-ms 1534438419097 \
  --log-url https://fake_url"