GitHub

Speed measurement

Here we will compare the speed of plotting UnfoldMakie with MNE (Python) and EEGLAB (MATLAB).

Three cases are measured:

  • Single topoplot
  • Topoplot series with 50 topoplots
  • Topoplott animation with 50 timestamps

Note that the results of benchmarking on your computer and on Github may differ.

using UnfoldMakie
using TopoPlots
using BenchmarkTools
using Observables
using CairoMakie
using Statistics
using PythonPlot;
using PyMNE;
    CondaPkg Found dependencies: /home/runner/.julia/packages/CondaPkg/8GjrP/CondaPkg.toml
    CondaPkg Found dependencies: /home/runner/.julia/packages/PyMNE/AlJE6/CondaPkg.toml
    CondaPkg Found dependencies: /home/runner/.julia/packages/PythonCall/83z4q/CondaPkg.toml
    CondaPkg Found dependencies: /home/runner/.julia/packages/PythonPlot/oS8x4/CondaPkg.toml
    CondaPkg Dependencies already up to date

Data input

dat, positions = TopoPlots.example_data()
df = UnfoldMakie.eeg_array_to_dataframe(dat[:, :, 1], string.(1:length(positions)));

Topoplots

UnfoldMakie.jl

@benchmark plot_topoplot(dat[:, 320, 1]; positions = positions)
BenchmarkTools.Trial: 71 samples with 1 evaluation per sample.
 Range (min … max):  46.941 ms …    1.319 s  ┊ GC (min … max):  0.00% … 95.96%
 Time  (median):     49.811 ms               ┊ GC (median):     0.00%
 Time  (mean ± σ):   75.526 ms ± 162.277 ms  ┊ GC (mean ± σ):  33.32% ± 15.51%

  █                                                             
  █▆▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▄ ▁
  46.9 ms       Histogram: log(frequency) by time       578 ms <

 Memory estimate: 11.81 MiB, allocs estimate: 194551.

UnfoldMakie.jl with DelaunayMesh

@benchmark plot_topoplot(
    dat[:, 320, 1];
    positions = positions,
    topo_interpolation = (; interpolation = DelaunayMesh()),
)
BenchmarkTools.Trial: 56 samples with 1 evaluation per sample.
 Range (min … max):  57.024 ms …    1.581 s  ┊ GC (min … max):  0.00% … 95.81%
 Time  (median):     62.555 ms               ┊ GC (median):     0.00%
 Time  (mean ± σ):   89.585 ms ± 202.883 ms  ┊ GC (mean ± σ):  30.19% ± 12.80%

                      ▄▁▁▁▄  ▁▁▁ █▄   ▄▁▄   ▁                   
  ▆▁▁▁▆▁▁▁▁▆▁▁▆▁▁▁▁▆▁▆█████▆▆███▆██▁▆▆███▁▁▆█▆▆▆▁▆▆▁▆▁▆▁▁▁▁▁▁▆ ▁
  57 ms           Histogram: frequency by time         67.6 ms <

 Memory estimate: 11.81 MiB, allocs estimate: 194558.

MNE

posmat = collect(reduce(hcat, [[p[1], p[2]] for p in positions])')
pypos = Py(posmat).to_numpy()
pydat = Py(dat[:, 320, 1])

@benchmark begin
    f = PythonPlot.figure()
    PyMNE.viz.plot_topomap(
        pydat,
        pypos,
        sphere = 1.1,
        extrapolate = "box",
        cmap = "RdBu_r",
        sensors = false,
        contours = 6,
    )
    f.show()
end
BenchmarkTools.Trial: 320 samples with 1 evaluation per sample.
 Range (min … max):  13.464 ms … 213.789 ms  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     14.166 ms               ┊ GC (median):    0.00%
 Time  (mean ± σ):   15.642 ms ±  15.966 ms  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▂▅▇█▇▅                                                        
  ███████▁▅▅▁▄▄▁▁▁▁▁▁▁▁▁▁▁▁▁▁▄▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▄ ▆
  13.5 ms       Histogram: log(frequency) by time      25.4 ms <

 Memory estimate: 3.30 KiB, allocs estimate: 98.

Topoplot series

Note that UnfoldMakie and MNE have different defaults for displaying topoplot series. UnfoldMakie in plot_topoplot averages over time samples. MNE in plot_topopmap displays single samples without averaging.

UnfoldMakie.jl

@benchmark begin
    plot_topoplotseries(
        df;
        bin_num = 50,
        positions = positions,
        axis = (; xlabel = "Time windows [s]"),
    )
end
BenchmarkTools.Trial: 2 samples with 1 evaluation per sample.
 Range (min … max):  2.363 s …    2.799 s  ┊ GC (min … max): 0.00% … 11.16%
 Time  (median):     2.581 s               ┊ GC (median):    6.05%
 Time  (mean ± σ):   2.581 s ± 308.237 ms  ┊ GC (mean ± σ):  6.05% ±  7.89%

  █                                                        █  
  █▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█ ▁
  2.36 s         Histogram: frequency by time          2.8 s <

 Memory estimate: 351.58 MiB, allocs estimate: 5076848.

MNE

easycap_montage = PyMNE.channels.make_standard_montage("standard_1020")
ch_names = pyconvert(Vector{String}, easycap_montage.ch_names)[1:64]
info = PyMNE.create_info(PyList(ch_names), ch_types = "eeg", sfreq = 1)
info.set_montage(easycap_montage)
simulated_epochs = PyMNE.EvokedArray(Py(dat[:, :, 1]), info)

@benchmark simulated_epochs.plot_topomap(1:50)
BenchmarkTools.Trial: 7 samples with 1 evaluation per sample.
 Range (min … max):  753.465 ms … 772.882 ms  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     764.776 ms               ┊ GC (median):    0.00%
 Time  (mean ± σ):   763.952 ms ±   8.257 ms  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▁       ▁   ▁                      ▁                ▁       █  
  █▁▁▁▁▁▁▁█▁▁▁█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▁▁▁▁▁▁▁█ ▁
  753 ms           Histogram: frequency by time          773 ms <

 Memory estimate: 2.36 KiB, allocs estimate: 68.

MATLAB

Running MATLAB on a GitHub Action is not easy. So we benchmarked three consecutive executions (on a screenshot) on a server with an AMD EPYC 7452 32-core processor. Note that Github and the server we used for MATLAB benchmarking are two different computers, which can give different timing results.

Animation

The main advantage of Julia is the speed with which the figures are updated.

timestamps = range(1, 50, step = 1)
framerate = 50
50

UnfoldMakie with .gif

vals = vec(dat[:, :, 1])
p01, p99 = quantile(vals, [0.01, 0.99])
m = max(abs(p01), abs(p99))
cr = Float32.((-m, m))
@benchmark begin
    f = Makie.Figure()
    dat_obs = Observable(dat[:, 1, 1])
    plot_topoplot!(f[1, 1], dat_obs, positions = positions, visual = (; contours = false, colorrange = cr),)
    record(f, "topoplot_animation_UM.gif", timestamps; framerate = framerate) do t
        dat_obs[] = @view(dat[:, t, 1])
    end
end
BenchmarkTools.Trial: 3 samples with 1 evaluation per sample.
 Range (min … max):  2.045 s …   2.229 s  ┊ GC (min … max): 0.00% … 1.97%
 Time  (median):     2.182 s              ┊ GC (median):    1.84%
 Time  (mean ± σ):   2.152 s ± 95.753 ms  ┊ GC (mean ± σ):  1.30% ± 1.10%

  █                                         █             █  
  █▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▁▁▁▁▁▁▁▁▁▁▁▁▁█ ▁
  2.04 s         Histogram: frequency by time        2.23 s <

 Memory estimate: 120.33 MiB, allocs estimate: 416721.

MNE with .gif

@benchmark begin
    fig, anim = simulated_epochs.animate_topomap(
        times = Py(timestamps),
        frame_rate = framerate,
        blit = false,
        image_interp = "cubic", # same as CloughTocher
    )
    anim.save("topomap_animation_mne.gif", writer = "ffmpeg", fps = framerate)
end
BenchmarkTools.Trial: 1 sample with 1 evaluation per sample.
 Single result which took 8.790 s (0.00% GC) to evaluate,
 with a memory estimate of 3.03 KiB, over 96 allocations.

Note, that due to some bugs in (probably) PythonCall topoplot is black and white.

This page was generated using Literate.jl.