Christopher Kormanyos
|
ee4748d087
|
In <boost/castfloat.h>, temporarily(?) de-activate __float128 == boost::float128_t.
|
2014-02-01 20:44:31 +01:00 |
|
Christopher Kormanyos
|
33813b7efe
|
In <boost/cstdfloat.hpp>, corrected static assertion on 64 digits in float80_t.
|
2014-01-30 20:59:23 +01:00 |
|
Christopher Kormanyos
|
ca054f9e78
|
In Bessel / Airy zeros, reliance on numeric_limits for non-specialized types corrected.
|
2014-01-29 22:38:13 +01:00 |
|
Christopher Kormanyos
|
eee4394266
|
In Bessel / Airy zeros, remove use of numeric_limits<> in root calculations.
|
2014-01-29 22:02:29 +01:00 |
|
Christopher Kormanyos
|
4ea3d12425
|
Removed the use of std::numeric_limits<> in Bessel / Airy zero calculations in favor of tools is boost::math::tools.
|
2014-01-29 21:54:50 +01:00 |
|
Christopher Kormanyos
|
ad420f7db8
|
Merge branch 'develop' of https://github.com/boostorg/math into develop
|
2014-01-28 20:29:33 +01:00 |
|
Christopher Kormanyos
|
dbbc33905e
|
In <cstdfloat.hpp>, correct the detection of float80_t for historical MSVC.
|
2014-01-28 20:28:39 +01:00 |
|
jzmaddock
|
1b84923f70
|
Make sure internal calls to expm1/log1p are properly qualified and pass a policy.
|
2014-01-24 11:59:59 +00:00 |
|
jzmaddock
|
9baf90f56a
|
Merge branch 'bessel-derivatives' into develop
|
2014-01-23 16:55:55 +00:00 |
|
Christopher Kormanyos
|
a3c1b273cb
|
Refine and correct the comments based on recent consensus(es) with co-authors.
|
2014-01-21 23:59:16 +01:00 |
|
jzmaddock
|
7613d0dd45
|
Rename Bessel derivatives to primes.
|
2014-01-21 19:08:14 +00:00 |
|
Christopher Kormanyos
|
82824159ed
|
Clean up cbrt_epsilon and <cstdfloat.hpp>.
|
2014-01-19 21:24:55 +01:00 |
|
Christopher Kormanyos
|
afcf566a53
|
Eliminate use of numeric_limits in <cstdfloat.hpp> for GCC's __float128.
|
2014-01-19 21:21:50 +01:00 |
|
Christopher Kormanyos
|
f9d81b179c
|
Correct a bug that I caused in cbrt_epsilon(). But I don't know why cbrt did not work?
|
2014-01-19 21:21:12 +01:00 |
|
jzmaddock
|
21d8d99e6f
|
Add concept checks for Bessel derivatives and fix up resulting errors.
|
2014-01-19 17:42:24 +00:00 |
|
jzmaddock
|
29a6f932f1
|
Add forward declarations for Bessel derivatives.
|
2014-01-19 16:45:28 +00:00 |
|
jzmaddock
|
d5099fc86a
|
Suppress some msvc warnings.
|
2014-01-19 16:44:49 +00:00 |
|
jzmaddock
|
7c82510c66
|
Merge https://github.com/AntonBikineev/bessel-derivatives into branch bessel-derivatives preserving history as best we can.
|
2014-01-19 13:14:12 +00:00 |
|
jzmaddock
|
431284519f
|
Move directory structure around to match what we need for the merge.
|
2014-01-19 13:06:47 +00:00 |
|
Christopher Kormanyos
|
c63676b055
|
Improve small argument handling in the generic tgamma(). Thereby add new function cbrt_epsilon() in tools/precision.hpp.
|
2014-01-19 13:10:31 +01:00 |
|
Christopher Kormanyos
|
cefd32709c
|
Initial commit of <boost/cstdfloat.hpp>.
|
2014-01-19 13:09:41 +01:00 |
|
Christopher Kormanyos
|
fcbd3d6e25
|
Move <boost/cstdfloat.hpp> to the proper directory.
|
2014-01-19 12:45:53 +01:00 |
|
Christopher Kormanyos
|
572562d271
|
Merge branch 'bernoulli-tgamma' of https://github.com/boostorg/math into bernoulli-tgamma
|
2014-01-19 11:24:20 +01:00 |
|
Christopher Kormanyos
|
5fc976c306
|
Move header <cstdfloat.hpp> to mainstream Boost.
|
2014-01-19 11:22:40 +01:00 |
|
Nikhar Agrawal
|
036945c71f
|
Added work done on polygamma function during GSoC
|
2014-01-19 15:21:34 +05:30 |
|
Christopher Kormanyos
|
0557af86a3
|
In <cstdfloat.hpp>, add support for the BOOST_NO_FLOAT128_T preprocessor definition.
|
2014-01-18 21:42:23 +01:00 |
|
Christopher Kormanyos
|
9c62c70dff
|
Remove float24_t from <cstdfloat.hpp>. Imrove the preliminary docs.
|
2014-01-18 15:18:55 +01:00 |
|
Christopher Kormanyos
|
3d61999cea
|
A few tweaks in <cstdfloat.hpp> and preliminary text-based docs.
|
2014-01-18 15:05:16 +01:00 |
|
Christopher Kormanyos
|
deb2daac04
|
In <cstdfloat.hpp> corrected the number of digits in float24_t.
|
2014-01-15 11:38:10 +01:00 |
|
Christopher Kormanyos
|
45a097e93b
|
To <cstdfloat.hpp>, add support for tiny floats float16_t and float24_t.
|
2014-01-14 22:40:24 +01:00 |
|
Christopher Kormanyos
|
93f0d2271c
|
Improve the comments in <ctsdfloat.hpp>.
|
2014-01-13 22:42:44 +01:00 |
|
Christopher Kormanyos
|
a8eff735fd
|
Re-correct the logic of un-defines in <cstdfloat.hpp>.
|
2014-01-13 20:21:33 +01:00 |
|
Christopher Kormanyos
|
cd6f694730
|
Improve clarity of <cstdfloat.hpp>.
|
2014-01-13 20:00:56 +01:00 |
|
Christopher Kormanyos
|
6855c35598
|
In <cstdfloat.hpp>, add support for __float128 from GCC's libquadmath.
|
2014-01-12 19:35:38 +01:00 |
|
Christopher Kormanyos
|
83ca3c282b
|
<cstdfloat> now stops at the least of float, double, or long double when querying <float.h>.
|
2014-01-12 15:52:51 +01:00 |
|
Christopher Kormanyos
|
d6bbde4a84
|
Clean up comments, etc. in preliminary <cstdfloat.hpp>.
|
2014-01-11 12:22:12 +01:00 |
|
Christopher Kormanyos
|
af02c1634e
|
Correct errors found via GCC for AVR tiny 8-bit microcontroller.
|
2014-01-10 23:20:00 +01:00 |
|
Christopher Kormanyos
|
cdc7b3b8cc
|
Improved floating-point type detection for a wider range of compilers.
|
2014-01-10 22:49:32 +01:00 |
|
Christopher Kormanyos
|
caf895fd23
|
Manual merge from bernoulli-numbers branch to bernoulli-tgamma branch.
|
2014-01-10 20:25:15 +01:00 |
|
Christopher Kormanyos
|
16efa7b842
|
Correct the min/max values in the macros.
|
2014-01-10 20:15:09 +01:00 |
|
Christopher Kormanyos
|
e373870c01
|
Preliminary concept implemntation of <cstdfloat>.
|
2014-01-10 19:37:33 +01:00 |
|
Christopher Kormanyos
|
c363d125f4
|
Initial commit of a very rudimentary sketch of <cstdfloat>.
|
2014-01-10 08:36:28 +01:00 |
|
Christopher Kormanyos
|
cd74da0ed1
|
Add tiny argument treatment to multiprecision tgamma(). Add tools cbrt_epsilon().
|
2014-01-10 08:35:43 +01:00 |
|
jzmaddock
|
1706a3c2b6
|
Improve error messages from overflow.
|
2014-01-09 10:17:01 +00:00 |
|
Christopher Kormanyos
|
9c6f195e4d
|
Implement cubed root of epsilon and preliminary multiprecision tgamma() and lgamma() for very small argument.
|
2014-01-08 22:58:11 +01:00 |
|
jzmaddock
|
c102fb62e9
|
Merge branch 'bernoulli-tgamma' into bernoulli-numbers
|
2014-01-07 13:06:14 +00:00 |
|
jzmaddock
|
49b30df548
|
Fix the order in which recursion occurs so that a) we don't end up dividing by zero if |z| < epsilon, and b) don't lose precision when z is close to epsilon.
|
2014-01-06 17:24:20 +00:00 |
|
Christopher Kormanyos
|
34ddaf71b5
|
In multiprecision lgamma() as well as tgamma() now use factorial for small unsigned integer arguments.
|
2014-01-05 21:23:33 +01:00 |
|
jzmaddock
|
5a4d1a85af
|
Forward to tabulated factorials when possible: fixes absurdities like lgamma(1) != 0 which occur otherwise.
|
2014-01-05 17:49:12 +00:00 |
|
Christopher Kormanyos
|
8aa1b7068e
|
Support multiprecision tgamma() and lgamma() for types without a specialization of numeric_limits.
|
2014-01-05 15:33:20 +01:00 |
|