| 7a2c3f0 | Paul Zimmermann | 23 November 2015, 15:47:15 UTC | this should work on 32-bit machines too | 23 November 2015, 15:47:15 UTC |
| 5dc6e9f | Paul Zimmermann | 23 November 2015, 15:36:21 UTC | added assert to detect GMP bug | 23 November 2015, 15:36:21 UTC |
| c821fbc | Paul Zimmermann | 23 November 2015, 15:10:18 UTC | error message should be the same in all cases | 23 November 2015, 15:10:18 UTC |
| 97e7e62 | Pierrick Gaudry | 23 November 2015, 12:51:09 UTC | missing \ in regexp | 23 November 2015, 12:51:26 UTC |
| 86b5f00 | Emmanuel Thomé | 23 November 2015, 12:40:41 UTC | Simplify cado-nfs-client daemonization. The double-fork pattern is useless if the grandparent does exit() straight away. At least so it seems. (the previous approach with SIGUSR2 was doomed because it was not checking the condition atomically after signal.pause(). signal.pause() is not pthread_cond_wait. The latter has a mutex, for a reason!). | 23 November 2015, 12:47:23 UTC |
| f9d6f84 | Paul Zimmermann | 23 November 2015, 12:40:55 UTC | fixed bug in purge when 0 <= excess < keep | 23 November 2015, 12:40:55 UTC |
| 5bf0bdb | Paul Zimmermann | 23 November 2015, 07:41:51 UTC | better sieving parameters for c190 | 23 November 2015, 07:41:51 UTC |
| 648519c | Paul Zimmermann | 20 November 2015, 14:54:44 UTC | use 1 thread by default in purge (instead of 4) | 20 November 2015, 14:54:44 UTC |
| fa7c004 | Paul Zimmermann | 20 November 2015, 12:42:51 UTC | removed nbmergemax too | 20 November 2015, 12:42:51 UTC |
| a853543 | Paul Zimmermann | 20 November 2015, 10:19:16 UTC | removed forbw, renamed coverNmax to target_density (increased to 170) | 20 November 2015, 10:19:16 UTC |
| 0a6a50c | Emmanuel Thomé | 19 November 2015, 23:50:54 UTC | now that sides are called side-0 and side-1, stop using sidename[side][0] | 19 November 2015, 23:51:00 UTC |
| 09ee2a1 | Paul Zimmermann | 19 November 2015, 18:06:00 UTC | now we can do: cado-nfs.py ... tasks.sieve.batch=True | 19 November 2015, 18:06:00 UTC |
| 13d7ecf | Paul Zimmermann | 19 November 2015, 17:58:06 UTC | changed verbose level in batch mode | 19 November 2015, 17:58:06 UTC |
| 28a6025 | Emmanuel Thomé | 19 November 2015, 09:33:48 UTC | typo | 19 November 2015, 09:33:48 UTC |
| 16aae64 | Emmanuel Thomé | 19 November 2015, 09:30:41 UTC | modified README | 19 November 2015, 09:31:03 UTC |
| dcebc4b | Cyril Bouvier | 19 November 2015, 09:25:20 UTC | Merge branch 'master' of git+ssh://scm.gforge.inria.fr/gitroot/cado-nfs/cado-nfs | 19 November 2015, 09:25:20 UTC |
| 08c4e4c | Paul Zimmermann | 19 November 2015, 08:51:47 UTC | updated supported platforms | 19 November 2015, 08:51:47 UTC |
| 909cc40 | Cyril Bouvier | 19 November 2015, 08:44:49 UTC | Merge branch 'master' of git+ssh://scm.gforge.inria.fr/gitroot/cado-nfs/cado-nfs | 19 November 2015, 08:44:49 UTC |
| 5bfeddc | Cyril Bouvier | 18 November 2015, 17:15:04 UTC | Fix bugs in check_rels. Adapt testcheck_rels The semantic of lpb is changed to behave like in las, freerel, characters | 18 November 2015, 17:15:04 UTC |
| 76e71db | Paul Zimmermann | 18 November 2015, 15:40:11 UTC | new item to do | 18 November 2015, 15:40:11 UTC |
| 949002e | Paul Zimmermann | 18 November 2015, 09:38:05 UTC | added a new test with batch mode | 18 November 2015, 09:38:05 UTC |
| 3bc032b | Paul Zimmermann | 18 November 2015, 07:30:14 UTC | oops | 18 November 2015, 07:30:14 UTC |
| caaad24 | Paul Zimmermann | 17 November 2015, 21:09:16 UTC | now the batch mode can be used simply as las ... -batch | 17 November 2015, 21:09:16 UTC |
| e7d8062 | Paul Zimmermann | 17 November 2015, 13:16:58 UTC | minor changes | 17 November 2015, 13:16:58 UTC |
| 0e38f70 | Paul Zimmermann | 17 November 2015, 13:15:49 UTC | increased maxlevel | 17 November 2015, 13:15:49 UTC |
| 6919e75 | Paul Zimmermann | 17 November 2015, 13:10:17 UTC | removed duplicate check | 17 November 2015, 13:10:17 UTC |
| a24113d | Paul Zimmermann | 17 November 2015, 08:55:23 UTC | put back default coverNmax to 100 for DL | 17 November 2015, 08:55:23 UTC |
| 41bcbe0 | Paul Zimmermann | 16 November 2015, 21:35:17 UTC | increased coverNmax from 100 to 150 | 16 November 2015, 21:35:17 UTC |
| 32827ae | Paul Zimmermann | 14 November 2015, 18:58:13 UTC | typo | 14 November 2015, 18:58:13 UTC |
| 0866ec3 | Emmanuel Thomé | 13 November 2015, 16:12:25 UTC | Fix bug with keep_rolling_checkpoints | 13 November 2015, 16:12:41 UTC |
| 2198de9 | Paul Zimmermann | 13 November 2015, 16:09:12 UTC | fix for write outside bound error | 13 November 2015, 16:09:12 UTC |
| e7d7b75 | Emmanuel Thomé | 13 November 2015, 15:04:38 UTC | commit 3de0815 was wrong. icc groks -x assembler. | 13 November 2015, 15:04:38 UTC |
| 2111ae4 | Paul Zimmermann | 13 November 2015, 14:44:07 UTC | icc 16 is ok | 13 November 2015, 14:44:07 UTC |
| a755991 | Paul Zimmermann | 13 November 2015, 14:39:58 UTC | oops in previous commit | 13 November 2015, 14:39:58 UTC |
| bf5b71f | Emmanuel Thomé | 13 November 2015, 13:18:49 UTC | fix gcc-isms in matops.cpp and binary-dotprod-backends.h | 13 November 2015, 14:34:00 UTC |
| a59d6a9 | Paul Zimmermann | 13 November 2015, 14:25:46 UTC | better SSE2 code, working with ICC too | 13 November 2015, 14:25:46 UTC |
| 813f9d6 | Paul Zimmermann | 13 November 2015, 13:05:04 UTC | with this patch, the README test works with icc 14.0.3 too | 13 November 2015, 13:05:04 UTC |
| 717b447 | Paul Zimmermann | 13 November 2015, 12:43:40 UTC | now SSE2 code in vaddmul_tiny_64K_64L works with icc too | 13 November 2015, 12:43:40 UTC |
| 3de0815 | Emmanuel Thomé | 13 November 2015, 10:59:39 UTC | do not try $CC -x assembler with non-gcc. | 13 November 2015, 10:59:59 UTC |
| f5c8eea | Emmanuel Thomé | 13 November 2015, 09:32:37 UTC | asm in bwc was mistakenly disabled | 13 November 2015, 10:59:59 UTC |
| cd14a5e | Paul Zimmermann | 13 November 2015, 10:48:30 UTC | only vaddmul_tiny_64K_64L seems to fail with icc | 13 November 2015, 10:48:30 UTC |
| 6f03efe | Paul Zimmermann | 13 November 2015, 09:11:30 UTC | patch to allow compilation without gas assembly | 13 November 2015, 09:11:30 UTC |
| 564dd06 | Paul Zimmermann | 13 November 2015, 08:36:38 UTC | removed macro with DISABLE_ASM (not defined/documented anywhere) | 13 November 2015, 08:36:38 UTC |
| 70fa4c4 | Paul Zimmermann | 12 November 2015, 20:27:41 UTC | another fix for icc version 14.0.3 (gcc version 4.4.7 compatibility) | 12 November 2015, 20:27:41 UTC |
| 214424d | Paul Zimmermann | 12 November 2015, 17:20:55 UTC | typo | 12 November 2015, 17:20:55 UTC |
| fe5d387 | Emmanuel Thomé | 12 November 2015, 15:38:49 UTC | Darwin has sysctl machdep.cpu with all sorts of useful stuff. | 12 November 2015, 15:38:49 UTC |
| cec716b | Emmanuel Thomé | 12 November 2015, 10:21:41 UTC | Oh I love escaping backslashes. 8 for the price of 1, isn't that cool ? | 12 November 2015, 10:21:41 UTC |
| 3bfd232 | Emmanuel Thomé | 12 November 2015, 10:19:45 UTC | supposedly more portable... http://bugs.python.org/issue8080 | 12 November 2015, 10:19:49 UTC |
| c5a0a18 | Emmanuel Thomé | 12 November 2015, 10:16:42 UTC | path woes | 12 November 2015, 10:16:42 UTC |
| d1981bf | Emmanuel Thomé | 12 November 2015, 10:11:19 UTC | more portable paths | 12 November 2015, 10:11:19 UTC |
| afbd8c8 | Emmanuel Thomé | 12 November 2015, 09:16:27 UTC | oops | 12 November 2015, 09:16:27 UTC |
| b97d434 | Emmanuel Thomé | 12 November 2015, 09:15:28 UTC | provide mac-compatible implementation of readlink -f | 12 November 2015, 09:15:28 UTC |
| 04812df | Emmanuel Thomé | 12 November 2015, 08:32:01 UTC | avoid symlinks in doctests | 12 November 2015, 08:32:01 UTC |
| a946304 | Paul Zimmermann | 11 November 2015, 21:11:17 UTC | added patchlevel to version number | 11 November 2015, 21:11:17 UTC |
| a1be25b | Pierrick Gaudry | 11 November 2015, 20:05:00 UTC | Threadpool is now based on a priority queue. This helps the scheduling of fill-in-buckets tasks: the ones that create many hits are scheduled first, to help having the buckets evenly filled in the end. | 11 November 2015, 20:05:19 UTC |
| 759a0ef | Emmanuel Thomé | 11 November 2015, 17:19:07 UTC | fix defaultlocale() returning None in some cases | 11 November 2015, 17:19:07 UTC |
| 91df37e | Pierrick Gaudry | 11 November 2015, 09:37:03 UTC | Hum... Previous commit was too optimistic. Put again margin in thread workspaces and give a better error message when buckets are full. | 11 November 2015, 09:37:03 UTC |
| 1d46ff9 | Pierrick Gaudry | 11 November 2015, 08:50:13 UTC | Memory seems to be solved? | 11 November 2015, 08:50:13 UTC |
| 04baaf8 | Pierrick Gaudry | 11 November 2015, 08:24:46 UTC | what did I have in mind??? | 11 November 2015, 08:24:46 UTC |
| 55e5c14 | Pierrick Gaudry | 10 November 2015, 17:07:28 UTC | Again parameter position... this is really a pain | 10 November 2015, 17:07:28 UTC |
| faaa7f4 | Pierrick Gaudry | 10 November 2015, 17:01:13 UTC | oops | 10 November 2015, 17:01:13 UTC |
| 086a507 | Pierrick Gaudry | 10 November 2015, 16:49:54 UTC | -dlp-no-keep is back. | 10 November 2015, 16:49:54 UTC |
| ab5d4df | Pierrick Gaudry | 10 November 2015, 16:31:05 UTC | wrong param in test_gfp3 | 10 November 2015, 16:31:05 UTC |
| 99cba5d | Pierrick Gaudry | 10 November 2015, 16:23:09 UTC | Fix gfp2 test. | 10 November 2015, 16:23:09 UTC |
| 53d04a8 | Pierrick Gaudry | 10 November 2015, 16:11:18 UTC | fix python toplevel test | 10 November 2015, 16:11:18 UTC |
| f476130 | Paul Zimmermann | 10 November 2015, 15:49:50 UTC | increased version to 2.2 | 10 November 2015, 15:49:50 UTC |
| fd5ff48 | Emmanuel Thomé | 10 November 2015, 15:11:16 UTC | mysterious | 10 November 2015, 15:11:16 UTC |
| 6698af7 | Pierrick Gaudry | 10 November 2015, 15:07:41 UTC | fix python_toplevel test | 10 November 2015, 15:07:41 UTC |
| af1b59a | Pierrick Gaudry | 10 November 2015, 14:58:19 UTC | Fix DLP in GF(p) with new main script. | 10 November 2015, 14:58:19 UTC |
| 9fedbca | Paul Zimmermann | 10 November 2015, 14:23:12 UTC | nchar is mandatory too | 10 November 2015, 14:23:12 UTC |
| e6d2320 | Paul Zimmermann | 10 November 2015, 14:04:20 UTC | m and n are mandatory for bwc (no default values) | 10 November 2015, 14:04:20 UTC |
| a8169ef | Emmanuel Thomé | 10 November 2015, 13:24:03 UTC | various little changes | 10 November 2015, 13:24:12 UTC |
| e75a337 | Emmanuel Thomé | 10 November 2015, 12:56:03 UTC | have command-line --slaves win over slaves.nrclients | 10 November 2015, 13:24:12 UTC |
| 83de57e | Paul Zimmermann | 10 November 2015, 13:15:27 UTC | update default installation prefix | 10 November 2015, 13:15:27 UTC |
| 288ca18 | Paul Zimmermann | 10 November 2015, 13:06:10 UTC | enf od review of main README | 10 November 2015, 13:06:10 UTC |
| 9a3e2f1 | Paul Zimmermann | 10 November 2015, 12:42:43 UTC | slaves.hostnames is mandatory to start clients | 10 November 2015, 12:42:43 UTC |
| 1826a28 | Paul Zimmermann | 10 November 2015, 12:07:23 UTC | update parameters.F9 | 10 November 2015, 12:07:23 UTC |
| 17d1d83 | Paul Zimmermann | 10 November 2015, 11:58:56 UTC | Merge branch 'master' of git+ssh://scm.gforge.inria.fr/git/cado-nfs/cado-nfs | 10 November 2015, 11:58:56 UTC |
| 5ab0695 | Paul Zimmermann | 10 November 2015, 11:55:26 UTC | review of main README (not finished) | 10 November 2015, 11:55:26 UTC |
| ead296e | Pierrick Gaudry | 10 November 2015, 10:54:21 UTC | more information. | 10 November 2015, 10:54:21 UTC |
| 798f62b | Paul Zimmermann | 10 November 2015, 08:57:41 UTC | added hwloc | 10 November 2015, 08:57:41 UTC |
| fbf9f68 | Paul Zimmermann | 10 November 2015, 08:56:48 UTC | Merge branch 'master' of git+ssh://scm.gforge.inria.fr/git/cado-nfs/cado-nfs | 10 November 2015, 08:56:48 UTC |
| db8c47e | Paul Zimmermann | 10 November 2015, 08:56:38 UTC | typo | 10 November 2015, 08:56:38 UTC |
| 07d7529 | Pierrick Gaudry | 10 November 2015, 08:50:50 UTC | up_path has no trailing / | 10 November 2015, 08:50:50 UTC |
| 5566e75 | Paul Zimmermann | 10 November 2015, 08:49:31 UTC | speedups | 10 November 2015, 08:49:31 UTC |
| 2ff9d4a | Paul Zimmermann | 10 November 2015, 08:19:40 UTC | improved documentation | 10 November 2015, 08:19:40 UTC |
| 26890db | Emmanuel Thomé | 09 November 2015, 20:10:18 UTC | one more path which should not be there. | 09 November 2015, 20:10:18 UTC |
| f0e97cb | Emmanuel Thomé | 09 November 2015, 20:08:26 UTC | python3.2 does not have os.uname().sysname. Use os.uname()[0], safer. | 09 November 2015, 20:08:26 UTC |
| 9b94380 | Emmanuel Thomé | 09 November 2015, 18:40:25 UTC | comment out parameters which really should not be in default parameter files | 09 November 2015, 18:40:25 UTC |
| 7a71417 | Emmanuel Thomé | 09 November 2015, 18:00:07 UTC | default parameter files should not set N | 09 November 2015, 18:00:07 UTC |
| 2a0fa64 | Emmanuel Thomé | 09 November 2015, 17:21:44 UTC | fedora17 with python 3.2 dislikes this, it is weird | 09 November 2015, 17:21:44 UTC |
| 656669c | Emmanuel Thomé | 09 November 2015, 17:18:16 UTC | python fix | 09 November 2015, 17:18:16 UTC |
| c0c1ab0 | Emmanuel Thomé | 09 November 2015, 16:34:34 UTC | fix doc | 09 November 2015, 16:40:32 UTC |
| 902237b | Emmanuel Thomé | 09 November 2015, 14:58:51 UTC | kill factor.sh ; make the toplevel scripts be cado-nfs.py and cado-nfs-client.py | 09 November 2015, 16:40:21 UTC |
| e4412bf | Cyril Bouvier | 09 November 2015, 12:16:59 UTC | dup1: fix some bugs with -n arguments (the bounds were never checked) | 09 November 2015, 12:16:59 UTC |
| 0007527 | Paul Zimmermann | 09 November 2015, 11:54:26 UTC | make matops.cpp compile with icc | 09 November 2015, 11:54:26 UTC |
| a336e8c | Paul Zimmermann | 09 November 2015, 11:03:29 UTC | workaround icpc internal error | 09 November 2015, 11:03:29 UTC |
| 4e50048 | Paul Zimmermann | 09 November 2015, 10:37:18 UTC | Revert "should be better for win7" This reverts commit 3a011701932829cc47e50ca2e9b2c90cbc176b40. | 09 November 2015, 10:37:18 UTC |
| 391886c | Paul Zimmermann | 09 November 2015, 09:24:46 UTC | oops | 09 November 2015, 09:24:46 UTC |
| 84c1fd8 | Paul Zimmermann | 09 November 2015, 09:22:28 UTC | fix issues with icc | 09 November 2015, 09:22:28 UTC |
