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

Comparing Different Compilers
PrevUpHomeNext

Comparing Different Compilers

By running our performance test suite multiple times, we can compare the effect of different compilers: as might be expected, the differences are generally small compared to say disabling internal use of long double. However, there are still gains to be main, particularly from some of the commercial offerings:

Table 23.12. Compiler Comparison on Windows x64

Function

Microsoft Visual C++ version 14.2
boost 1.73

GNU C++ version 9.2.0
boost 1.73

GNU C++ version 9.2.0
boost 1.73
promote_double<false>

assoc_laguerre

1.41
(179ns)

1.08
(137ns)

1.00
(127ns)

assoc_legendre

1.76
(248ns)

1.36
(192ns)

1.00
(141ns)

beta

1.00
(123ns)

2.62
(322ns)

1.93
(237ns)

beta (incomplete)

1.00
(470ns)

2.95
(1385ns)

1.58
(741ns)

cbrt

3.40
(51ns)

4.67
(70ns)

1.00
(15ns)

cyl_bessel_i

1.00
(281ns)

3.38
(949ns)

1.38
(387ns)

cyl_bessel_i (integer order)

1.00
(195ns)

3.06
(597ns)

1.00
(195ns)

cyl_bessel_j

1.00
(371ns)

2.39
(886ns)

1.35
(499ns)

cyl_bessel_j (integer order)

1.28
(123ns)

1.92
(184ns)

1.00
(96ns)

cyl_bessel_k

1.11
(385ns)

19.68
(6847ns)

1.00
(348ns)

cyl_bessel_k (integer order)

1.06
(217ns)

18.17
(3724ns)

1.00
(205ns)

cyl_neumann

1.17
(6696ns)

1.76
(10032ns)

1.00
(5715ns)

cyl_neumann (integer order)

1.00
(158ns)

2.20
(348ns)

1.59
(252ns)

digamma

1.00
(20ns)

3.45
(69ns)

2.30
(46ns)

ellint_1

1.57
(390ns)

1.41
(349ns)

1.00
(248ns)

ellint_1 (complete)

1.64
(77ns)

1.64
(77ns)

1.00
(47ns)

ellint_2

1.81
(702ns)

1.50
(583ns)

1.00
(388ns)

ellint_2 (complete)

3.11
(84ns)

2.11
(57ns)

1.00
(27ns)

ellint_3

3.47
(1381ns)

1.68
(670ns)

1.00
(398ns)

ellint_3 (complete)

inf
(802ns)

-nan(ind)
(0ns)

-nan(ind)
(0ns)

ellint_rc

1.55
(59ns)

2.21
(84ns)

1.00
(38ns)

ellint_rd

1.32
(271ns)

1.26
(260ns)

1.00
(206ns)

ellint_rf

1.27
(62ns)

1.94
(95ns)

1.00
(49ns)

ellint_rj

1.46
(264ns)

2.29
(414ns)

1.00
(181ns)

erf

1.30
(43ns)

1.85
(61ns)

1.00
(33ns)

erfc

1.06
(54ns)

1.76
(90ns)

1.00
(51ns)

expint

1.00
(27ns)

3.41
(92ns)

2.22
(60ns)

expint (En)

1.00
(106ns)

1.94
(206ns)

1.29
(137ns)

expm1

1.00
(11ns)

3.00
(33ns)

2.36
(26ns)

gamma_p

1.00
(303ns)

2.00
(605ns)

1.17
(355ns)

gamma_p_inv

1.00
(1266ns)

1.85
(2341ns)

1.15
(1460ns)

gamma_q

1.00
(294ns)

2.10
(618ns)

1.21
(356ns)

gamma_q_inv

1.00
(1194ns)

1.66
(1987ns)

1.14
(1357ns)

ibeta

1.00
(512ns)

2.63
(1344ns)

1.31
(673ns)

ibeta_inv

1.00
(1910ns)

2.49
(4751ns)

1.48
(2822ns)

ibetac

1.00
(525ns)

2.60
(1365ns)

1.27
(668ns)

ibetac_inv

1.00
(1676ns)

2.85
(4778ns)

1.74
(2910ns)

jacobi_cn

1.00
(181ns)

3.10
(561ns)

2.00
(362ns)

jacobi_dn

1.00
(203ns)

3.03
(616ns)

1.93
(392ns)

jacobi_sn

1.00
(202ns)

2.81
(568ns)

1.73
(350ns)

laguerre

1.02
(107ns)

1.07
(112ns)

1.00
(105ns)

legendre

1.11
(283ns)

1.25
(320ns)

1.00
(255ns)

legendre Q

1.00
(309ns)

1.51
(466ns)

1.15
(354ns)

lgamma

1.00
(80ns)

2.67
(214ns)

2.00
(160ns)

log1p

1.00
(14ns)

2.07
(29ns)

1.21
(17ns)

polygamma

1.00
(4193ns)

1.85
(7743ns)

1.91
(8018ns)

sph_bessel

1.01
(668ns)

1.48
(975ns)

1.00
(661ns)

sph_neumann

1.07
(1138ns)

2.96
(3153ns)

1.00
(1064ns)

tgamma

1.00
(74ns)

3.50
(259ns)

2.14
(158ns)

tgamma (incomplete)

1.00
(208ns)

2.30
(478ns)

1.64
(342ns)

trigamma

1.00
(12ns)

2.83
(34ns)

1.17
(14ns)

zeta

1.00
(117ns)

2.65
(310ns)

1.89
(221ns)


Table 23.13. Compiler Comparison on linux

Function

GNU C++ version 9.2.1 20191008
boost 1.73

GNU C++ version 9.2.1 20191008
boost 1.73
promote_double<false>

Clang version 9.0.0 (tags/RELEASE_900/final)
boost 1.73

Clang version 9.0.0 (tags/RELEASE_900/final)
boost 1.73
promote_double<false>

Intel C++ C++0x mode version 1910
boost 1.73

Intel C++ C++0x mode version 1910
boost 1.73
promote_double<false>

assoc_laguerre

1.17
(156ns)

1.08
(143ns)

1.72
(229ns)

1.31
(174ns)

1.15
(153ns)

1.00
(133ns)

assoc_legendre

2.91
(279ns)

1.78
(171ns)

2.86
(275ns)

1.00
(96ns)

2.79
(268ns)

2.16
(207ns)

beta

10.52
(789ns)

1.40
(105ns)

10.56
(792ns)

1.73
(130ns)

6.07
(455ns)

1.00
(75ns)

beta (incomplete)

4.98
(2076ns)

1.26
(524ns)

5.44
(2267ns)

1.00
(417ns)

3.15
(1315ns)

1.16
(484ns)

cbrt

2.52
(58ns)

1.09
(25ns)

2.35
(54ns)

1.22
(28ns)

1.91
(44ns)

1.00
(23ns)

cyl_bessel_i

5.05
(1182ns)

1.02
(239ns)

10.73
(2511ns)

1.74
(406ns)

3.99
(933ns)

1.00
(234ns)

cyl_bessel_i (integer order)

4.41
(767ns)

1.00
(174ns)

5.09
(886ns)

1.03
(179ns)

3.63
(631ns)

1.12
(195ns)

cyl_bessel_j

3.80
(1067ns)

1.28
(359ns)

4.72
(1327ns)

1.55
(436ns)

3.05
(857ns)

1.00
(281ns)

cyl_bessel_j (integer order)

3.10
(319ns)

1.05
(108ns)

4.73
(487ns)

1.00
(103ns)

2.30
(237ns)

1.12
(115ns)

cyl_bessel_k

10.37
(3962ns)

1.00
(382ns)

11.55
(4414ns)

1.52
(582ns)

12.41
(4739ns)

1.09
(416ns)

cyl_bessel_k (integer order)

19.23
(3289ns)

1.06
(181ns)

19.78
(3382ns)

1.27
(217ns)

20.02
(3423ns)

1.00
(171ns)

cyl_neumann

2.18
(8826ns)

1.19
(4814ns)

3.23
(13059ns)

1.31
(5275ns)

2.35
(9503ns)

1.00
(4041ns)

cyl_neumann (integer order)

3.67
(466ns)

1.35
(172ns)

4.49
(570ns)

1.15
(146ns)

2.87
(364ns)

1.00
(127ns)

digamma

2.30
(62ns)

1.30
(35ns)

3.11
(84ns)

1.33
(36ns)

1.85
(50ns)

1.00
(27ns)

ellint_1

2.26
(310ns)

1.12
(154ns)

6.96
(954ns)

1.25
(171ns)

3.22
(441ns)

1.00
(137ns)

ellint_1 (complete)

4.65
(93ns)

3.15
(63ns)

6.00
(120ns)

1.15
(23ns)

1.90
(38ns)

1.00
(20ns)

ellint_2

3.81
(704ns)

1.28
(236ns)

8.95
(1655ns)

1.36
(251ns)

4.36
(807ns)

1.00
(185ns)

ellint_2 (complete)

1.83
(44ns)

1.04
(25ns)

6.25
(150ns)

1.25
(30ns)

2.00
(48ns)

1.00
(24ns)

ellint_3

3.27
(609ns)

1.17
(217ns)

5.37
(998ns)

1.32
(245ns)

2.84
(528ns)

1.00
(186ns)

ellint_3 (complete)

-nan
(0ns)

-nan
(0ns)

-nan
(0ns)

-nan
(0ns)

-nan
(0ns)

-nan
(0ns)

ellint_rc

2.75
(66ns)

1.29
(31ns)

4.71
(113ns)

1.42
(34ns)

1.83
(44ns)

1.00
(24ns)

ellint_rd

4.37
(428ns)

1.31
(128ns)

9.24
(906ns)

1.31
(128ns)

4.17
(409ns)

1.00
(98ns)

ellint_rf

2.47
(84ns)

1.44
(49ns)

4.62
(157ns)

1.32
(45ns)

1.85
(63ns)

1.00
(34ns)

ellint_rj

4.14
(426ns)

1.36
(140ns)

7.18
(740ns)

1.31
(135ns)

3.33
(343ns)

1.00
(103ns)

erf

1.69
(54ns)

1.28
(41ns)

2.56
(82ns)

1.38
(44ns)

1.34
(43ns)

1.00
(32ns)

erfc

1.65
(89ns)

1.20
(65ns)

2.46
(133ns)

1.26
(68ns)

1.22
(66ns)

1.00
(54ns)

expint

2.93
(85ns)

1.45
(42ns)

2.90
(84ns)

1.31
(38ns)

2.21
(64ns)

1.00
(29ns)

expint (En)

2.42
(242ns)

1.14
(114ns)

2.76
(276ns)

1.19
(119ns)

1.84
(184ns)

1.00
(100ns)

expm1

1.90
(19ns)

1.60
(16ns)

1.50
(15ns)

1.40
(14ns)

1.80
(18ns)

1.00
(10ns)

gamma_p

2.43
(618ns)

1.02
(260ns)

2.95
(749ns)

1.00
(254ns)

2.15
(545ns)

1.03
(262ns)

gamma_p_inv

2.37
(2328ns)

1.06
(1038ns)

2.73
(2676ns)

1.06
(1045ns)

1.92
(1883ns)

1.00
(982ns)

gamma_q

2.49
(629ns)

1.00
(253ns)

3.28
(831ns)

1.11
(280ns)

2.18
(551ns)

1.04
(263ns)

gamma_q_inv

2.73
(2016ns)

1.30
(958ns)

3.22
(2381ns)

1.30
(962ns)

1.73
(1275ns)

1.00
(739ns)

ibeta

4.56
(1979ns)

1.21
(524ns)

4.60
(1996ns)

1.00
(434ns)

3.35
(1453ns)

1.13
(491ns)

ibeta_inv

3.22
(5552ns)

1.10
(1902ns)

3.49
(6013ns)

1.03
(1781ns)

2.58
(4437ns)

1.00
(1722ns)

ibetac

4.34
(2008ns)

1.27
(590ns)

4.34
(2008ns)

1.00
(463ns)

3.11
(1438ns)

1.17
(543ns)

ibetac_inv

3.32
(5533ns)

1.11
(1853ns)

3.55
(5915ns)

1.03
(1722ns)

2.48
(4131ns)

1.00
(1666ns)

jacobi_cn

3.74
(483ns)

1.26
(163ns)

4.98
(643ns)

1.61
(208ns)

2.67
(345ns)

1.00
(129ns)

jacobi_dn

3.64
(503ns)

1.25
(173ns)

4.70
(648ns)

1.49
(206ns)

2.72
(375ns)

1.00
(138ns)

jacobi_sn

4.99
(614ns)

1.46
(180ns)

5.20
(640ns)

1.72
(212ns)

2.57
(316ns)

1.00
(123ns)

laguerre

1.11
(92ns)

1.08
(90ns)

1.41
(117ns)

1.11
(92ns)

1.17
(97ns)

1.00
(83ns)

legendre

1.12
(282ns)

1.00
(251ns)

1.21
(304ns)

1.30
(326ns)

1.39
(349ns)

1.08
(270ns)

legendre Q

1.24
(390ns)

1.00
(314ns)

1.29
(404ns)

1.14
(358ns)

1.42
(445ns)

1.11
(347ns)

lgamma

2.65
(265ns)

1.17
(117ns)

3.86
(386ns)

1.24
(124ns)

1.90
(190ns)

1.00
(100ns)

log1p

1.11
(20ns)

1.22
(22ns)

1.06
(19ns)

1.00
(18ns)

1.06
(19ns)

1.00
(18ns)

polygamma

16.70
(8883ns)

1.42
(756ns)

6.56
(3491ns)

1.00
(532ns)

7.49
(3985ns)

1.05
(559ns)

sph_bessel

2.36
(1122ns)

1.24
(588ns)

3.31
(1574ns)

1.41
(673ns)

2.12
(1010ns)

1.00
(476ns)

sph_neumann

3.40
(2842ns)

1.21
(1015ns)

3.59
(3009ns)

1.21
(1012ns)

3.14
(2627ns)

1.00
(837ns)

tgamma

4.72
(472ns)

1.44
(144ns)

5.70
(570ns)

1.05
(105ns)

2.29
(229ns)

1.00
(100ns)

tgamma (incomplete)

3.28
(669ns)

1.13
(231ns)

3.88
(792ns)

1.14
(232ns)

2.21
(451ns)

1.00
(204ns)

trigamma

2.08
(25ns)

1.00
(12ns)

4.00
(48ns)

1.50
(18ns)

2.00
(24ns)

1.08
(13ns)

zeta

4.21
(455ns)

1.00
(108ns)

5.77
(623ns)

1.20
(130ns)

2.98
(322ns)

1.08
(117ns)


Copyright © 2006-2021 Nikhar Agrawal, Anton Bikineev, Matthew Borland, Paul A. Bristow, Marco Guazzone, Christopher Kormanyos, Hubert Holin, Bruno Lalande, John Maddock, Evan Miller, Jeremy Murphy, Matthew Pulver, Johan Råde, Gautam Sewani, Benjamin Sobotta, Nicholas Thompson, Thijs van den Berg, Daryle Walker and Xiaogang Zhang

Distributed under the Boost Software License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)

PrevUpHomeNext