William Hart
|
c50bf32dfc
|
Added netburst tuning.
|
2014-03-21 15:05:00 +00:00 |
|
BrianGladman
|
74524a3e29
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-03-20 16:16:06 +00:00 |
|
William Hart
|
a307e2221f
|
New penryn tuning.
|
2014-03-20 16:12:00 +00:00 |
|
BrianGladman
|
8a93786668
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-03-20 15:53:33 +00:00 |
|
William Hart
|
3b6a11e4c5
|
Added tuning code from GMP for gcd, gcdext, jacobi and fac_ui.
|
2014-03-20 15:28:01 +00:00 |
|
BrianGladman
|
c3a91f5403
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-03-20 00:06:26 +00:00 |
|
William Hart
|
e6b6751c97
|
Update tuning for penryn.
|
2014-03-19 23:41:58 +00:00 |
|
William Hart
|
471fddeece
|
Added tuning for small version of sb_divappr_q and sb_div_qr.
|
2014-03-19 23:09:38 +00:00 |
|
William Hart
|
5d35ea55a4
|
Better k10 tuning values.
|
2014-03-19 17:48:14 +00:00 |
|
William Hart
|
4fa350f51d
|
Hack FFT usage into mpn_mulmod_2expp1.
|
2014-03-19 17:32:01 +00:00 |
|
BrianGladman
|
f17b419e36
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-02-28 17:10:31 +00:00 |
|
William Hart
|
8e3504e372
|
Updated to new bin_ui and bin_uiui code from GMP 5.1.3. Added a version of
mpn_divrem_hensel_rsh_qr_1 that takes a precomputed inverse.
|
2014-02-28 16:44:53 +00:00 |
|
BrianGladman
|
6e53dff494
|
add new factorial code and tests to Visual Studio builds
|
2014-02-28 15:12:11 +00:00 |
|
BrianGladman
|
edcefb28cb
|
Update Windows Visual Studio 2013 build for latest changes
|
2014-02-27 22:16:06 +00:00 |
|
William Hart
|
67809f39a2
|
Update to GMP 5.1.3 factorial code.
|
2014-02-27 21:35:04 +00:00 |
|
William Hart
|
68c4a8d583
|
Apparently I didn't update these correctly.
|
2014-02-27 17:35:55 +00:00 |
|
William Hart
|
3c20c75b3b
|
Whoops, forgot to git add these files at the last commit.
|
2014-02-27 16:11:37 +00:00 |
|
William Hart
|
0c96bff79f
|
Add rudimentary mpn_mulmod_bnm1. Update to new gcd and xgcd code from GMP.
|
2014-02-27 16:00:08 +00:00 |
|
William Hart
|
d4adcd7ce6
|
Some very minor speedups/cleanups.
|
2014-02-27 12:30:53 +00:00 |
|
William Hart
|
c7cdb70393
|
Small improvement to sb_div_q.
|
2014-02-25 15:58:30 +00:00 |
|
BrianGladman
|
94c011b8a3
|
1. add revised core 2 assembler to the Windows build
2. add revised nehalem tuning on Windows
3. fix bug in speed on Windows x64
|
2014-02-24 21:07:18 +00:00 |
|
BrianGladman
|
e6047c1c9e
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-02-23 21:30:13 +00:00 |
|
William Hart
|
b65ad160ab
|
Fix slowdown on penryn, retune k10.
|
2014-02-23 17:47:35 +00:00 |
|
William Hart
|
796fb5b77e
|
Switch back to single function for sb_divappr_q.
|
2014-02-21 16:15:43 +00:00 |
|
William Hart
|
d5a444bd02
|
Use GMP add/submul_1 on core2 as well.
|
2014-02-21 15:25:41 +00:00 |
|
William Hart
|
faa82ed6c9
|
Update penryn tuning.
|
2014-02-21 14:27:14 +00:00 |
|
William Hart
|
738fb46982
|
Fix a bug in sb_divappr_q.
|
2014-02-21 14:14:41 +00:00 |
|
William Hart
|
e9cd157e50
|
Improve tail of sb_divappr_q and fix bug.
|
2014-02-21 13:41:37 +00:00 |
|
William Hart
|
e41d78e02b
|
Switch to GMP add/submul_1.asm on Penryn.
|
2014-02-21 11:22:14 +00:00 |
|
William Hart
|
8bb084e6d2
|
Speed up sb_divappr_q.
|
2014-02-21 10:57:56 +00:00 |
|
William Hart
|
5548023724
|
Switch to using both types of precomputed inverse.
|
2014-02-20 20:21:04 +00:00 |
|
BrianGladman
|
71f91234c2
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-02-19 07:38:40 +00:00 |
|
William Hart
|
fc7b94378a
|
Updated atom tuning values.
|
2014-02-18 22:42:48 +00:00 |
|
BrianGladman
|
2e9c90554c
|
update windows builds for GMP sqr_basecase change
|
2014-02-18 17:36:29 +00:00 |
|
William Hart
|
8873ab5738
|
Merge remote-tracking branch 'brian/master'
|
2014-02-18 17:24:53 +00:00 |
|
BrianGladman
|
1a41ad921f
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2014-02-18 16:25:37 +00:00 |
|
William Hart
|
a1a268dc7e
|
Corrected nehalem tuning value.
|
2014-02-18 12:42:07 +00:00 |
|
William Hart
|
c8c2a49c43
|
Update nehalem tuning values.
|
2014-02-18 11:29:26 +00:00 |
|
William Hart
|
2a741c97d0
|
On nehalem the GMP sqr_basecase is faster. Thus both nehalem and westmere can
use the mpn/x86_64/sqr_basecase.asm file.
|
2014-02-18 11:25:37 +00:00 |
|
William Hart
|
d10ddc5184
|
Adjust atom kara_sqr_n crossover.
|
2014-02-18 11:15:46 +00:00 |
|
BrianGladman
|
d76d8284b9
|
add GMP sqr_basecase assembler code for Windows
|
2014-02-17 23:06:06 +00:00 |
|
BrianGladman
|
b261394905
|
add new sqr_basecase code on Windows
|
2014-02-17 23:02:51 +00:00 |
|
William Hart
|
7c3794b221
|
Updated some tuning values for penryn.
|
2014-02-17 21:08:46 +00:00 |
|
William Hart
|
581be69339
|
Update sqr fft crossover.
|
2014-02-17 20:40:31 +00:00 |
|
William Hart
|
c63be9f5f7
|
Update k102 tuning.
|
2014-02-17 20:38:58 +00:00 |
|
William Hart
|
e2f5dd4c02
|
Switched to using correct basecase threshold in kara_sqr_n.
|
2014-02-17 20:23:26 +00:00 |
|
William Hart
|
48544b2e95
|
Sandybridge is faster with GMP sqr_basecase.
|
2014-02-17 19:52:48 +00:00 |
|
William Hart
|
edb937d0e3
|
GMP sqr_basecase is faster on westmere.
|
2014-02-17 19:37:58 +00:00 |
|
William Hart
|
06b0fea603
|
Update tuning values for netburst.
|
2014-02-17 19:14:15 +00:00 |
|
William Hart
|
01fc86a25e
|
The GMP sqr_basecase is *way* faster on netburst.
|
2014-02-17 19:08:27 +00:00 |
|
William Hart
|
24d3b12a0a
|
Use GMP version of sqr_basecase on AMD K8/K10/etc and Intel Core2/Penryn/etc.
|
2014-02-17 16:43:45 +00:00 |
|
William Hart
|
d9416e6bee
|
Added some GMP m4 macros for new sqr_basecase code + license changes.
|
2014-02-17 15:54:07 +00:00 |
|
William Hart
|
449976bd8d
|
Fix a slowdown in FFT (actually in mulmod_2expp1_basecase) when squaring.
|
2014-02-17 15:48:41 +00:00 |
|
William Hart
|
41674225e1
|
Cleaned up lots of C code, adding lots of whitespace.
|
2014-02-14 23:25:39 +00:00 |
|
William Hart
|
8437e1f92a
|
Added comments to add_err1_n/add_err2_n.asm and fixed some comments in
sub_err1_n/sub_err2_n.asm.
|
2014-02-07 14:55:59 +00:00 |
|
William Hart
|
66827c7e00
|
Comments in sub_err1_n/sub_err2_n.asm.
|
2014-02-07 13:05:55 +00:00 |
|
William Hart
|
7e5589a8bb
|
Ran autoreconf -i to change autotools versions to those of development machine.
|
2014-02-04 13:40:51 +00:00 |
|
William Hart
|
fbf04dda95
|
Minor improvement of divide and conquer divapprox based on proof in paper.
|
2014-02-01 00:06:00 +00:00 |
|
Brian Gladman
|
d6ca8cf09a
|
Merge remote-tracking branch 'remotes/Bill/master'
|
2013-08-06 13:25:33 +01:00 |
|
Brian Gladman
|
fd76ae5284
|
adjust Intel macros in dedicated windows assembler
|
2013-08-06 13:25:11 +01:00 |
|
William Hart
|
0550f973bd
|
Merge remote-tracking branch 'brian/master'
|
2013-08-06 13:20:09 +01:00 |
|
Jean-Pierre Flori
|
b3438467e6
|
Ran "autoreconf -fiv --no-recursive".
|
2013-08-05 00:34:35 +02:00 |
|
Jean-Pierre Flori
|
4491e7b7da
|
Modify PROLOGUE declarations in x86_64w.
This let configure define correct HAVE_NATIVE_* constants.
This may break the VS builds.
|
2013-08-05 00:17:37 +02:00 |
|
Jean-Pierre Flori
|
bdd21aa9c4
|
Harmonize PROLOGUE in Win64 assembly files.
This breaks non-generic builds with MinGW.
|
2013-08-04 23:45:38 +02:00 |
|
Jean-Pierre Flori
|
f9c0c1381d
|
Fix for MPIR_AS.
|
2013-08-02 21:01:04 +02:00 |
|
Jean-Pierre Flori
|
40b0d3bf3a
|
Properly quote define so that m4 does not fail.
|
2013-08-01 19:07:47 +02:00 |
|
Jean-Pierre Flori
|
a2800f0920
|
Fix typos in assembly file names.
|
2013-08-01 18:59:45 +02:00 |
|
Jean-Pierre Flori
|
d222af9e8c
|
Ran "autoreconf -fi" and put back correct config.(guess|sub) files.
|
2013-07-31 14:51:32 +02:00 |
|
Brian Gladman
|
2b412ed5ab
|
change the define INTEL_COMPILER to __INTEL_COMPILER
|
2013-04-18 23:45:38 +01:00 |
|
Brian Gladman
|
a81efe8936
|
Update the C++ header file (mpirxx.h) to match GMP
Replace the incorrect INTEL_COMPILER define to __INTEL_COMPILER
|
2013-04-17 13:54:43 +01:00 |
|
Brian Gladman
|
95bdc34c19
|
add mingw/mingw64 inline assembler to Windows longlong_inc.h files
|
2013-04-16 16:44:11 +01:00 |
|
Brian Gladman
|
f85231351f
|
minor cosmetic changes to the Windows build files
|
2013-04-16 10:18:17 +01:00 |
|
Brian Gladman
|
4033c0636a
|
Merge branch 'master' of https://github.com/wbhart/mpir
|
2013-04-16 09:11:37 +01:00 |
|
William Hart
|
e225faebfc
|
Added comment.
|
2013-04-15 19:38:29 +01:00 |
|
William Hart
|
e050c73be6
|
Removed some function call overhead.
|
2013-04-15 19:22:10 +01:00 |
|
William Hart
|
89acc6b8ce
|
Switch all precomputed inverses to new ones.
|
2013-04-15 17:39:36 +01:00 |
|
William Hart
|
ae7c8a30b2
|
Added dc_divappr_q and fixed sb_divappr_q, general code clean up.
|
2013-04-15 17:16:38 +01:00 |
|
William Hart
|
f72389e1a1
|
Switched to new divrem32 macro and did some minor housekeeping.
|
2013-04-13 02:19:15 +01:00 |
|
William Hart
|
db729465d7
|
Added new sb_div_q code, updated some comments and copyright notices.
|
2013-04-12 21:49:49 +01:00 |
|
William Hart
|
0b033b07f2
|
Fast sb_divappr_q code.
|
2013-04-12 17:49:29 +01:00 |
|
William Hart
|
019ddbfb99
|
Removed mullow_n_basecase.asm, as the interface is useless.
|
2013-04-11 01:56:11 +01:00 |
|
William Hart
|
e0922e5ce5
|
Changed the name of a macro.
|
2013-04-11 01:43:58 +01:00 |
|
William Hart
|
44d27e44eb
|
Updated copyright notices.
|
2013-04-11 01:28:46 +01:00 |
|
William Hart
|
188955ab2a
|
Removed a couple of duplicated macros.
|
2013-04-11 01:24:45 +01:00 |
|
William Hart
|
e967731486
|
Move divrem fixup into macro.
|
2013-04-11 01:05:33 +01:00 |
|
William Hart
|
7b7ffec89c
|
Switched to new divrem basecase.
|
2013-04-11 00:18:00 +01:00 |
|
William Hart
|
8655d3b7e0
|
Tried a new kind of basecase division.
|
2013-04-09 03:20:58 +01:00 |
|
William Hart
|
ebe4f759e1
|
Actually, we call it mullow_n_basecase, so move the file yet again, and
correct comment in mullow_n_basecase.c.
|
2013-04-06 01:21:39 +01:00 |
|
William Hart
|
9e8c5a0278
|
Renamed mullo_basecase to mullow_basecase so it will be picked up.
|
2013-04-06 00:52:58 +01:00 |
|
William Hart
|
98c0d8d14c
|
Made m4 macros in mullo_basecase.asm resolve by adding macros from GMP to
x86_64-defs.m4.
|
2013-04-06 00:36:53 +01:00 |
|
William Hart
|
954857eedb
|
Initial commit of mullo_basecase.asm, obtained from the GNU MP project (5.1.1)
m4 macros JUMPTABSECT, JMPENT, FUNC_ENTRY, FUNC_EXIT probably don't expand
yet.
|
2013-04-06 00:08:00 +01:00 |
|
Brian Gladman
|
b5c32bf68d
|
change FFT threshold in mul.c
|
2013-03-09 21:15:06 +00:00 |
|
William Hart
|
f608007fc5
|
Line endings
|
2012-11-25 22:33:07 +00:00 |
|
wbhart
|
96290ef587
|
Copied final tuning params across from *nix.
|
2012-10-25 22:25:37 +00:00 |
|
wbhart
|
a05e0123fe
|
ARM and AMD Bobcat tuning values courtesty of Julien Puydt and Leif Lionhardy
respectively.
|
2012-10-25 22:14:23 +00:00 |
|
gladman
|
ce643bc361
|
move new *nix tuning values into windows
|
2012-10-25 15:39:46 +00:00 |
|
wbhart
|
a881af8585
|
Sandy Bridge tuning values courtesy of JP Flori.
|
2012-10-25 13:13:41 +00:00 |
|
gladman
|
76944a04f8
|
move new *nix tuning values into windows
|
2012-10-24 07:01:53 +00:00 |
|
wbhart
|
1a010f75a6
|
Lots of tuning values.
|
2012-10-24 01:58:08 +00:00 |
|
wbhart
|
70fc2bb3a0
|
(i)fft_blah => mpir_(i)fft_blah, n_revbin => mpir_revbin, fermat_to_mpz =>
mpir_fermat_to_mpz, random_fermat => mpir_random_fermat. Moved fft protos
to gmp-impl.h.
|
2012-10-18 16:29:08 +00:00 |
|