Boost C++ Libraries

...one of the most highly regarded and expertly designed C++ library projects in the world. Herb Sutter and Andrei Alexandrescu, C++ Coding Standards

This is the documentation for an old version of boost. Click here for the latest Boost documentation.
PrevUpHomeNext

Comparisons to Other Open Source Libraries

We've run our performance tests both for our own code, and against other open source implementations of the same functions. The results are presented below to give you a rough idea of how they all compare.

[Caution] Caution

You should exercise extreme caution when interpreting these results, relative performance may vary by platform, the tests use data that gives good code coverage of our code, but which may skew the results towards the corner cases. Finally, remember that different libraries make different choices with regard to performance verses numerical stability.

Comparison to GSL-1.9 and Cephes

All the results were measured on a 2.8GHz Intel Pentium 4, 2Gb RAM, Windows XP machine, with all the libraries compiled with Microsoft Visual C++ 2005 using the /Ox /arch:SSE2 options.

Function

Boost

GSL-1.9

Cephes

tgamma

1.50

(2.566e-007s)

1.54

(2.627e-007s)

1.00

(1.709e-007s)

lgamma

1.73

(2.688e-007s)

3.61

(5.621e-007s)

1.00

(1.556e-007s)

gamma_p and gamma_q

1.00

(9.504e-007s)

2.15

(2.042e-006s)

2.57

(2.439e-006s)

gamma_p_inv and gamma_q_inv

1.00

(3.631e-006s)

N/A

+INF [1]

ibeta and ibetac

1.00

(1.852e-006s)

1.07

(1.974e-006s)

1.07

(1.974e-006s)

ibeta_inv and ibetac_inv

1.00

(7.311e-006s)

N/A

2.24

(1.637e-005s)

[1] Cephes gets stuck in an infinite loop while trying to execute our test cases.

Comparison to the R Statistical Library on Windows

All the results were measured on a 2.8GHz Intel Pentium 4, 2Gb RAM, Windows XP machine, with the test program compiled with Microsoft Visual C++ 2005, and R-2.5.0 compiled in "standalone mode" with MinGW-3.4 (R-2.5.0 appears not to be buildable with Visual C++).

Table 43. A Comparison to the R Statistical Library on Windows XP

Statistical Function

Boost

R

Beta Distribution CDF

1.20

(1.916e-006s)

1.00

(1.597e-006s)

Beta Distribution Quantile

1.00

(6.570e-006s)

74.66 [1]

(4.905e-004s)

Binomial Distribution CDF

1.00

(5.276e-007s)

2.45

(1.293e-006s)

Binomial Distribution Quantile

1.00

(4.013e-006s)

1.32

(5.280e-006s)

Cauchy Distribution CDF

1.00

(1.231e-007s)

1.28

(1.576e-007s)

Cauchy Distribution Quantile

1.00

(1.498e-007s)

1.00

(1.498e-007s)

Chi Squared Distribution CDF

1.00

(7.889e-007s)

2.48

(1.955e-006s)

Chi Squared Distribution Quantile

1.00

(4.303e-006s)

1.61

(6.925e-006s)

Exponential Distribution CDF

1.00

(1.955e-007s)

1.97

(3.844e-007s)

Exponential Distribution Quantile

1.07

(1.206e-007s)

1.00

(1.126e-007s)

Fisher F Distribution CDF

1.00

(1.309e-006s)

2.12

(2.780e-006s)

Fisher F Distribution Quantile

1.00

(7.204e-006s)

1.78

(1.280e-005s)

Gamma Distribution CDF

1.00

(1.076e-006s)

2.07

(2.227e-006s)

Gamma Distribution Quantile

1.00

(5.189e-006s)

1.14

(5.937e-006s)

Log-normal Distribution CDF

1.00

(2.078e-007s)

1.41

(2.930e-007s)

Log-normal Distribution Quantile

1.00

(6.692e-007s)

1.63

(1.090e-006s)

Negative Binomial Distribution CDF

1.00

(9.005e-007s)

2.42

(2.178e-006s)

Negative Binomial Distribution Quantile

1.00

(9.601e-006s)

53.59 [2]

(5.145e-004s)

Normal Distribution CDF

1.00

(5.926e-008s)

3.01

(1.785e-007s)

Normal Distribution Quantile

1.00

(1.248e-007s)

1.05

(1.311e-007s)

Poisson Distribution CDF

1.00

(8.999e-007s)

2.42

(2.175e-006s)

Poisson Distribution

1.00

(1.853e-006s)

2.17

(4.014e-006s)

Students t Distribution CDF

1.00

(1.223e-006s)

1.13

(1.376e-006s)

Students t Distribution Quantile

1.00

(2.570e-006s)

1.04

(2.668e-006s)

Weibull Distribution CDF

1.00

(4.741e-007s)

1.46

(6.943e-007s)

Weibull Distribution Quantile

1.00

(7.926e-007s)

1.08

(8.542e-007s)

[1] There are a small number of our test cases where the R library fails to converge on a result: these tend to dominate the performance result.

[2] The R library appears to use a linear-search strategy, that can perform very badly in a small number of pathological cases, but may or may not be more efficient in "typical" cases


Comparison to the R Statistical Library on Linux

All the results were measured on a 2.8GHz Intel Pentium 4, 2Gb RAM, Mandriva Linux machine, with the test program and R-2.5.0 compiled with GNU G++ 4.2.0.

Table 44. A Comparison to the R Statistical Library on Linux

Statistical Function

Boost

R

Beta Distribution CDF

1.71

(3.508e-006s)

1.00

(2.050e-006s)

Beta Distribution Quantile

1.00

(1.294e-005s)

44.06 [1]

(5.701e-004s)

Binomial Distribution CDF

1.22

(1.342e-006s)

1.00

(1.104e-006s)

Binomial Distribution Quantile

1.36

(7.083e-006s)

1.00

(5.194e-006s)

Cauchy Distribution CDF

1.00

(1.372e-007s)

1.47

(2.017e-007s)

Cauchy Distribution Quantile

1.00

(1.542e-007s)

1.14

(1.752e-007s)

Chi Squared Distribution CDF

1.04

(1.820e-006s)

1.00

(1.753e-006s)

Chi Squared Distribution Quantile

1.39

(9.345e-006s)

1.00

(6.728e-006s)

Exponential Distribution CDF

1.00

(2.195e-007s)

1.17

(2.561e-007s)

Exponential Distribution Quantile

1.00

(1.123e-007s)

1.03

(1.155e-007s)

Fisher F Distribution CDF

1.00

(2.744e-006s)

1.08

(2.970e-006s)

Fisher F Distribution Quantile

1.14

(1.550e-005s)

1.00

(1.359e-005s)

Gamma Distribution CDF

1.29

(2.578e-006s)

1.00

(1.992e-006s)

Gamma Distribution Quantile

1.77

(1.020e-005s)

1.00

(5.757e-006s)

Log-normal Distribution CDF

1.00

(1.782e-007s)

2.00

(3.564e-007s)

Log-normal Distribution Quantile

1.00

(7.093e-007s)

1.07

(7.607e-007s)

Negative Binomial Distribution CDF

1.03

(2.209e-006s)

1.00

(2.141e-006s)

Negative Binomial Distribution Quantile

1.00

(1.826e-005s)

30.07 [2]

(5.490e-004s)

Normal Distribution CDF

1.00

(8.542e-008s)

2.09

(1.782e-007s)

Normal Distribution Quantile

1.00

(1.362e-007s)

1.26

(1.722e-007s)

Poisson Distribution CDF

1.10

(1.953e-006s)

1.00

(1.775e-006s)

Poisson Distribution

1.12

(4.214e-006s)

1.00

(3.752e-006s)

Students t Distribution CDF

1.55

(2.441e-006s)

1.00

(1.576e-006s)

Students t Distribution Quantile

1.33

(3.972e-006s)

1.00

(2.990e-006s)

Weibull Distribution CDF

1.00

(6.640e-007s)

1.06

(7.031e-007s)

Weibull Distribution Quantile

1.00

(7.504e-007s)

1.03

(7.710e-007s)

[1] There are a small number of our test cases where the R library fails to converge on a result: these tend to dominate the performance result.

[2] The R library appears to use a linear-search strategy, that can perform very badly in a small number of pathological cases, but may or may not be more efficient in "typical" cases



PrevUpHomeNext