Benchmarks

To get maximal performance, it is recommended to:

Rebuild the Julia system image for your CPU architecture
Pass -O3 and --check-bounds=no as command line flags to julia.

Run perf/runbenchmarks.jl to see benchmark results for the Atlas robot (v5) in the following scenarios:

Compute the joint-space mass matrix.
Do inverse dynamics.
Do forward dynamics.

Note that results on Travis builds are not at all representative because of code coverage. Results on a recent, fast machine with version 0.3.0:

Output of versioninfo():

Julia Version 0.6.0
Commit 9036443 (2017-06-19 13:05 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: Intel(R) Core(TM) i7-6950X CPU @ 3.00GHz
  WORD_SIZE: 64
  BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Haswell)
  LAPACK: libopenblas64_
  LIBM: libopenlibm
  LLVM: libLLVM-3.9.1 (ORCJIT, broadwell)

Mass matrix:

  memory estimate:  0 bytes
  allocs estimate:  0
  --------------
  minimum time:     13.790 μs (0.00% GC)
  median time:      14.263 μs (0.00% GC)
  mean time:        14.340 μs (0.00% GC)
  maximum time:     71.598 μs (0.00% GC)

Inverse dynamics:

  memory estimate:  0 bytes
  allocs estimate:  0
  --------------
  minimum time:     12.491 μs (0.00% GC)
  median time:      12.870 μs (0.00% GC)
  mean time:        60.109 μs (0.00% GC)
  maximum time:     471.387 ms (0.00% GC)

Forward dynamics:

  memory estimate:  64 bytes
  allocs estimate:  3
  --------------
  minimum time:     47.291 μs (0.00% GC)
  median time:      50.349 μs (0.00% GC)
  mean time:        153.675 μs (0.00% GC)
  maximum time:     1.033 s (0.00% GC)