2
0
mirror of https://github.com/boostorg/math.git synced 2026-01-29 19:52:08 +00:00
Commit Graph

627 Commits

Author SHA1 Message Date
Paul A. Bristow
95d3d4d406 Corrected support minimum to 0 (was 1).
[SVN r62380]
2010-06-02 14:49:22 +00:00
John Maddock
ac3a1ef164 Change to use new symbol visibility macros.
[SVN r62200]
2010-05-25 14:45:54 +00:00
John Maddock
200b0f7717 Fix cbrt's internal promotion rules.
Update tests to match.

[SVN r62087]
2010-05-18 17:24:01 +00:00
John Maddock
7ab7d21eae Fix failures when used with an expression-template enabled number type such as the gmpxx clases.
Add additional concept check for integer code using gmp classes.
Fixes #4139.

[SVN r61837]
2010-05-07 15:36:09 +00:00
John Maddock
c7e0965ce9 Previous fix doesn't work for later GCC versions, revert.
[SVN r61815]
2010-05-06 12:33:40 +00:00
John Maddock
0dd75b6c47 Prevent multiple definitions of BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS.
[SVN r61807]
2010-05-06 10:23:56 +00:00
John Maddock
45db115f9d Disable long double support when the C++ runtime does not support it either.
Fixes #3825.

[SVN r61806]
2010-05-06 10:21:18 +00:00
Douglas Gregor
81c8795897 Eliminate redundant (and incorrect) using declarations at function scope.
[SVN r61700]
2010-04-30 07:10:56 +00:00
Douglas Gregor
9257a12d1b Eliminate a redundant BOOST_MATH_STD_USING; we aren't allowed to redeclare declarations in function scope
[SVN r61595]
2010-04-26 22:57:02 +00:00
John Maddock
7a7e393485 Fix boost::math::round to handle halfway cases correctly.
Update tests accordingly.
Update boost::math::fmod docs to make sign and rounding direction clear.

[SVN r61252]
2010-04-13 16:50:46 +00:00
John Maddock
cd64856ff5 Improve the performance of the Bessel functions, and update docs.
[SVN r59274]
2010-01-27 13:16:14 +00:00
John Maddock
ea35733550 Add tgamma1pm1 speed test.
Suppress some more gcc warnings: in particular make sure that our C99 functions have the same throw specification as the functions in math.h.

[SVN r59129]
2010-01-18 17:07:03 +00:00
John Maddock
1a5044497d Change cbrt implementation to use a better performing algorithm.
Fix a few warnings along the way.

[SVN r59095]
2010-01-17 17:28:34 +00:00
John Maddock
488701a7e3 Suppress unused parameter warning.
[SVN r57509]
2009-11-09 12:26:02 +00:00
John Maddock
75f47f2cad remove surplus const-qualifiers.
[SVN r57354]
2009-11-04 12:15:05 +00:00
John Maddock
ba698c2297 Improve ibeta error handling, update and regenerate docs. Update tests.
Fixes #3533.

[SVN r57045]
2009-10-21 15:53:28 +00:00
John Maddock
cc52a19fb4 Improve ibeta error handling, and add new tests.
[SVN r57040]
2009-10-21 12:27:38 +00:00
John Maddock
ebc86f03d8 Update performance test results.
Adjust msvc performance tuning.

[SVN r56818]
2009-10-14 12:04:07 +00:00
John Maddock
21cd5f0d9f Update workarounds for Sun's latest compiler release.
[SVN r56711]
2009-10-11 12:20:21 +00:00
John Maddock
581f74e09b Fix duplicate definitions when BOOST_NO_EXPLICIT_FUNCTION_TEMPLATE_ARGUMENTS is defined.
[SVN r56679]
2009-10-09 16:52:27 +00:00
K. Noel Belcourt
fe4001165c Add PGI macro to get proper alignment.
[SVN r56653]
2009-10-08 17:52:10 +00:00
John Maddock
799b88be52 Change iteration limits to use get_epsilon().
[SVN r56650]
2009-10-08 16:03:25 +00:00
John Maddock
2e07572faf Add a check on iterations when Halley iterating.
Correct igamma inverse initial calculations, and adjust digits needed for convergence.
Fix erf calculation limits.
Adjust non-convergence behaviour in Halley iteration.
For issue #3408.

[SVN r56598]
2009-10-05 17:36:56 +00:00
John Maddock
71940bec79 Another round of performance tweaks for issue #3408.
These should make our igamma implementation comparable in performance to the dcdflib FORTRAN routine - at least as far as MSVC is concerned.

[SVN r56503]
2009-10-01 16:03:42 +00:00
John Maddock
8418e9d68e A few more minor performance tweaks for issue #3407.
[SVN r56455]
2009-09-28 17:06:39 +00:00
John Maddock
056d083a2f Changes relating to issue #3408.
Add hooks for the dcdflib to the incomplete gamma tests.
Add hooks for the dcdflib to the igamma performance tests.
Some small performance enhancements.

[SVN r56370]
2009-09-24 11:23:52 +00:00
John Maddock
f99607da5d Fix lookup ambiguity.
Fixes #3207.

[SVN r54284]
2009-06-23 16:39:50 +00:00
John Maddock
6feccbee3d Added a couple of static assertions to prevent misuse.
[SVN r53338]
2009-05-28 08:52:05 +00:00
Jeremiah Willcock
99b1ee34f7 Fixed most tab and min/max issues from trunk inspection report
[SVN r53141]
2009-05-20 19:19:00 +00:00
John Maddock
b36f2d6bb6 Fix for no long double math functions.
[SVN r52522]
2009-04-21 08:43:36 +00:00
John Maddock
0b017bc74c Fixes #2961.
GCC-4.0 workaround.

[SVN r52513]
2009-04-20 15:59:58 +00:00
John Maddock
bedc5b058a Add more instrumentation code, along with some AMD64/Linux fixes.
[SVN r52335]
2009-04-11 15:26:59 +00:00
John Maddock
3d45185864 Sort circuit evaluation when the result will be 0.
[SVN r52287]
2009-04-09 09:10:45 +00:00
John Maddock
97049ddaa7 Fix use of C99 macro names.
[SVN r51951]
2009-03-24 10:05:46 +00:00
John Maddock
9c124ea11c Fix bug in cyl_bessel_i that hits when v=0.5 and x is small.
Fix return type of signbit to match C99 std.
Update and regenerate docs.
Fixes #2877.

[SVN r51890]
2009-03-21 18:44:20 +00:00
John Maddock
a935d5e1cd Updated the sign functions to use Johan Rade's fp-utilities code.
Added tests for the sign functions.
Added docs for the sign functions and moved some sections around.

[SVN r51793]
2009-03-16 13:25:22 +00:00
John Maddock
e38209bdc0 Add instrumentation code and some FPU control options.
[SVN r51755]
2009-03-13 17:56:28 +00:00
John Maddock
f6a4cbad2e Misc. small platform specific fixes and expected error rate adjustments.
[SVN r51648]
2009-03-08 10:39:29 +00:00
John Maddock
f10e098a4f Adjust fudge-factors yet again.
[SVN r51578]
2009-03-03 18:18:17 +00:00
John Maddock
2246b77395 Change tests to link to external Boost.Test library as this speeds build times by several factors when building with Intel C++.
Change fpclassify to use our code rather than std::fpclassify with Intel C++ as it doesn't always compile the latter.

[SVN r51576]
2009-03-03 17:16:56 +00:00
John Maddock
cf84efb70b Adjust fudge factor upwards to try and fix some of the current regression test failures.
Add option to disable use of std::fpclassify.

[SVN r51501]
2009-03-01 12:19:04 +00:00
John Maddock
fef466be54 remove spurious euro sign in comment.
[SVN r51474]
2009-02-27 17:06:21 +00:00
John Maddock
323a108149 Merges changes from Sandbox version:
MPFR Support.
New distributions: laplace, logistic and hypergeometric.
Added PCH support to the Jamfiles for faster build times.
Added C# usage example.
Updated and regenerated docs.
Also fixes #2799 and updates the tests to catch this failure in future.

[SVN r51455]
2009-02-26 17:34:52 +00:00
Paul A. Bristow
cc7e9904cb Added reference to Wolfram in comments.
[SVN r51137]
2009-02-09 14:51:25 +00:00
John Maddock
5fad227b2a Fix bug in bessel_jn for n == -1.
Add new test case.
Checked that the other Bessel functions do not have the same issue.
Checked that real-valued -1 argument is fixed OK as well as integer argument.
Fixes #2733.

[SVN r51059]
2009-02-06 17:26:30 +00:00
John Maddock
34049f0624 Fix unused parameter warnings.
[SVN r50940]
2009-02-01 10:16:20 +00:00
John Maddock
840bbe720f Improve performance of non-central distributions and the logistic quantile.
Updated performance results.
Fixed a few bugs in performance testing code.
Added short-cut macros for new distributions in math.qbk.

[SVN r50896]
2009-01-30 11:00:12 +00:00
John Maddock
e595c6b837 Fix min/max usage.
[SVN r50895]
2009-01-30 10:48:17 +00:00
Paul A. Bristow
0b8d45eb88 New copyright date.
[SVN r50738]
2009-01-23 09:46:27 +00:00
John Maddock
0901fbb44b Fix remaining quantile tests with the hypergeometric.
[SVN r50723]
2009-01-22 10:01:54 +00:00