Skip to content

Run the benchmark executable

Ryan Newton edited this page May 23, 2014 · 3 revisions

Once you have a run-benchmark executable built, typing ./run-benchmark -h should show you a help screen like this:

USAGE: [set ENV VARS] run-benchmarks-racket-typing.exe [CMDLN OPTS]

Note: "CMDLN OPTS" includes patterns that select which benchmarks
     to run, based on name.

 Command Line Options:
      --no-recomp                Don't perform any compilation of benchmark executables.  Implies -no-clean.
      --no-clean                 Do not clean pre-existing executables before beginning.
      --shortrun                 Elide command line args to benchmarks to perform a testing rather than benchmarking run.
  -k  --keepgoing                Keep executing even after a build or run fails (default false)
      --with-cabal-install=PATH  Set the version of cabal-install to use for the cabal BuildMethod.
      --with-ghc=PATH            Set the path of the ghc compiler for the ghc BuildMethod.
      --trials=NUM               The number of times to run each benchmark.
      --runid=NUM                Force run ID to be a specific string; useful for completing failed runs
      --buildid=STR              Set the build ID used by the continuous integration system.
      --skipto=NUM               Skip ahead to a specific point in the configuration space.
  -h  --help                     Show this help message and exit.
  -V  --version                  Show the version and exit
      --name=NAME                Name for created/discovered table in the backend.

Fusion Table Options:
    --fusion-upload[=TABLEID]  enable fusion table upload.  Optionally set TABLEID; otherwise create/discover it.
    --clientid=ID              Use (and cache) Google client ID
    --clientsecret=STR         Use (and cache) Google client secret
    --fusion-test              Test authentication and list tables if possible.

Dribble plugin loaded.
  No additional flags, but uses --name for the base filename.

 Note: This bench harness was built against hsbencher library version 1.8.0.2

Note that plugin-specific sections will vary. Backend plugins are statically compiled into the benchmark executable (as determined by run-benchmarks.cabal and run-benchmarks.hs). In the future we might support dynamic loading of plugins, but we don't currently (e.g. v1.8).

The simplest run

./run-benchmarks.sh

A run with Fusion table upload

After you Create your own Fusion table, then you are ready to upload data to that table.