search

haskell-numerics / random-fu

A suite of Haskell libraries for representing, manipulating, and sampling random variables
42 stars 21 forks source link

Revert perf regression #59

Closed idontgetoutmuch closed 4 years ago

idontgetoutmuch commented 4 years ago

Sadly this https://github.com/haskell-numerics/random-fu/pull/45 caused a significant regression and this reverts it.

Test Name Current Mean Mean with PR Change (%)
dists/PureMT/uniform/Double/single sample 3.34E-08 3.68E-07 -90.91%
dists/PureMT/uniform/Double/sum of samples (implicit rvar) 2.05E-03 2.32E-02 -91.19%
dists/PureMT/uniform/Double/sum of samples (explicit rvar) 2.25E-03 2.35E-02 -90.42%
dists/PureMT/uniform/Double/sample of sum 3.25E-03 2.55E-02 -87.25%
dists/PureMT/uniform/Double/array of samples 2.57E-03 2.43E-02 -89.39%
dists/PureMT/uniform/Double/RVarT IO arrays 5.03E-03 2.90E-02 -82.63%
dists/PureMT/uniform/Int/single sample 2.38E-07 4.87E-07 -51.07%
dists/PureMT/uniform/Int/sum of samples (implicit rvar) 1.48E-02 3.05E-02 -51.56%
dists/PureMT/uniform/Int/sum of samples (explicit rvar) 1.50E-02 3.04E-02 -50.68%
dists/PureMT/uniform/Int/sample of sum 1.47E-02 2.96E-02 -50.36%
dists/PureMT/uniform/Int/array of samples 2.18E-02 3.61E-02 -39.43%
dists/PureMT/uniform/Int/RVarT IO arrays 1.77E-02 3.33E-02 -46.97%
dists/PureMT/stdUniform/Double/single sample 2.68E-08 4.21E-07 -93.63%
dists/PureMT/stdUniform/Double/sum of samples (implicit rvar) 1.54E-03 2.33E-02 -93.39%
dists/PureMT/stdUniform/Double/sum of samples (explicit rvar) 1.55E-03 2.48E-02 -93.77%
dists/PureMT/stdUniform/Double/sample of sum 2.66E-03 2.67E-02 -90.05%
dists/PureMT/stdUniform/Double/array of samples 1.99E-03 2.63E-02 -92.44%
dists/PureMT/stdUniform/Double/RVarT IO arrays 4.72E-03 3.25E-02 -85.46%
dists/PureMT/stdUniform/Int/single sample 2.52E-08 7.12E-07 -96.45%
dists/PureMT/stdUniform/Int/sum of samples (implicit rvar) 1.42E-03 4.34E-02 -96.73%
dists/PureMT/stdUniform/Int/sum of samples (explicit rvar) 1.34E-03 4.18E-02 -96.79%
dists/PureMT/stdUniform/Int/sample of sum 2.67E-03 5.12E-02 -94.78%
dists/PureMT/stdUniform/Int/array of samples 1.81E-03 4.33E-02 -95.82%
dists/PureMT/stdUniform/Int/RVarT IO arrays 4.35E-03 4.98E-02 -91.27%
dists/PureMT/poisson/Double/single sample 3.64E-07 1.97E-06 -81.50%
dists/PureMT/poisson/Double/sum of samples (implicit rvar) 2.32E-02 0.123280996 -81.17%
dists/PureMT/poisson/Double/sum of samples (explicit rvar) 2.35E-02 0.124861048 -81.20%
dists/PureMT/poisson/Double/sample of sum 2.41E-02 0.123237969 -80.44%
dists/PureMT/poisson/Double/array of samples 3.32E-02 0.129579289 -74.37%
dists/PureMT/poisson/Double/RVarT IO arrays 2.64E-02 0.119814662 -78.00%
dists/PureMT/poisson/Int/single sample 3.88E-07 1.82E-06 -78.72%
dists/PureMT/poisson/Int/sum of samples (implicit rvar) 2.47E-02 0.11587482 -78.70%
dists/PureMT/poisson/Int/sum of samples (explicit rvar) 2.30E-02 0.114998965 -80.02%
dists/PureMT/poisson/Int/sample of sum 2.29E-02 0.12210896 -81.27%
dists/PureMT/poisson/Int/array of samples 3.25E-02 0.127312738 -74.47%
dists/PureMT/poisson/Int/RVarT IO arrays 2.52E-02 0.11960705 -78.93%
dists/PureMT/binomial 10/Double/single sample 6.59E-07 3.99E-06 -83.46%
dists/PureMT/binomial 10/Double/sum of samples (implicit rvar) 4.31E-02 0.258988272 -83.37%
dists/PureMT/binomial 10/Double/sum of samples (explicit rvar) 4.44E-02 0.255071938 -82.58%
dists/PureMT/binomial 10/Double/sample of sum 0.411518783 0.670825859 -38.65%
dists/PureMT/binomial 10/Double/array of samples 0.101623275 0.326977502 -68.92%
dists/PureMT/binomial 10/Double/RVarT IO arrays 5.95E-02 0.270118585 -77.97%
dists/PureMT/binomial 10/Int/single sample 6.61E-07 4.01E-06 -83.54%
dists/PureMT/binomial 10/Int/sum of samples (implicit rvar) 4.30E-02 0.256739649 -83.25%
dists/PureMT/binomial 10/Int/sum of samples (explicit rvar) 4.33E-02 0.257709118 -83.22%
dists/PureMT/binomial 10/Int/sample of sum 0.292883399 0.501134391 -41.56%
dists/PureMT/binomial 10/Int/array of samples 9.02E-02 0.316737005 -71.52%
dists/PureMT/binomial 10/Int/RVarT IO arrays 8.17E-02 0.302595815 -73.01%
dists/PureMT/stdNormal/single sample 7.10E-08 6.86E-07 -89.65%
dists/PureMT/stdNormal/sum of samples (implicit rvar) 4.41E-03 4.38E-02 -89.93%
dists/PureMT/stdNormal/sum of samples (explicit rvar) 4.45E-03 4.40E-02 -89.89%
dists/PureMT/stdNormal/sample of sum 5.15E-03 4.54E-02 -88.67%
dists/PureMT/stdNormal/array of samples 6.73E-03 4.73E-02 -85.77%
dists/PureMT/stdNormal/RVarT IO arrays 7.15E-03 4.74E-02 -84.89%
dists/PureMT/exponential/single sample 3.81E-08 3.69E-07 -89.66%
dists/PureMT/exponential/sum of samples (implicit rvar) 2.20E-03 2.33E-02 -90.55%
dists/PureMT/exponential/sum of samples (explicit rvar) 2.25E-03 2.37E-02 -90.47%
dists/PureMT/exponential/sample of sum 5.80E-03 2.88E-02 -79.88%
dists/PureMT/exponential/array of samples 2.76E-03 2.43E-02 -88.62%
dists/PureMT/exponential/RVarT IO arrays 7.46E-03 3.10E-02 -75.96%
dists/PureMT/beta/single sample 4.20E-07 2.53E-06 -83.41%
dists/PureMT/beta/sum of samples (implicit rvar) 2.65E-02 0.161357275 -83.55%
dists/PureMT/beta/sum of samples (explicit rvar) 2.69E-02 0.165472193 -83.75%
dists/PureMT/beta/sample of sum 2.79E-02 0.164298111 -83.01%
dists/PureMT/beta/array of samples 4.01E-02 0.175814198 -77.17%
dists/PureMT/beta/RVarT IO arrays 3.34E-02 0.167400184 -80.03%
dists/PureMT/gamma/single sample 1.48E-07 1.16E-06 -87.22%
dists/PureMT/gamma/sum of samples (implicit rvar) 8.16E-03 7.35E-02 -88.90%
dists/PureMT/gamma/sum of samples (explicit rvar) 8.16E-03 7.35E-02 -88.90%
dists/PureMT/gamma/sample of sum 9.40E-03 7.45E-02 -87.38%
dists/PureMT/gamma/array of samples 1.86E-02 8.59E-02 -78.28%
dists/PureMT/gamma/RVarT IO arrays 1.14E-02 7.83E-02 -85.38%
dists/PureMT/triangular/single sample 1.24E-07 4.88E-07 -74.54%
dists/PureMT/triangular/sum of samples (implicit rvar) 7.90E-03 3.19E-02 -75.26%
dists/PureMT/triangular/sum of samples (explicit rvar) 7.90E-03 3.16E-02 -74.98%
dists/PureMT/triangular/sample of sum 8.54E-03 3.23E-02 -73.54%
dists/PureMT/triangular/array of samples 1.54E-02 3.95E-02 -60.93%
dists/PureMT/triangular/RVarT IO arrays 1.06E-02 3.53E-02 -70.01%
dists/PureMT/rayleigh/single sample 1.32E-07 5.00E-07 -73.70%
dists/PureMT/rayleigh/sum of samples (implicit rvar) 8.60E-03 3.19E-02 -73.04%
dists/PureMT/rayleigh/sum of samples (explicit rvar) 8.38E-03 3.19E-02 -73.75%
dists/PureMT/rayleigh/sample of sum 9.19E-03 3.32E-02 -72.31%
dists/PureMT/rayleigh/array of samples 1.34E-02 3.83E-02 -65.06%
dists/PureMT/rayleigh/RVarT IO arrays 1.11E-02 3.56E-02 -68.90%
dists/PureMT/dirichlet 2.88E-02 9.68E-02 -70.22%
dists/PureMT/multinomial/many p/small n 9.45E-03 3.41E-02 -72.31%
dists/PureMT/multinomial/many p/medium n 9.05E-02 0.328863485 -72.49%
dists/PureMT/multinomial/many p/large n 0.205577328 0.730400811 -71.85%
dists/PureMT/multinomial/few p/small n 8.36E-03 3.38E-02 -75.29%
dists/PureMT/multinomial/few p/medium n 8.60E-02 0.306485231 -71.94%
dists/PureMT/multinomial/few p/large n 0.190808324 0.680886071 -71.98%
dists/PureMT/shuffle 0.108362626 0.127708992 -15.15%