9586224 | Emmanuel Thomé | 31 January 2021, 22:58:47 UTC | Merge branch 'fix_assert_expensive' into 'master' Fix assert expensive See merge request cado-nfs/cado-nfs!21 | 31 January 2021, 22:58:47 UTC |
58e68e4 | Alexander Kruppa | 31 January 2021, 22:22:37 UTC | Avoid errors with ASSERT_EXPENSIVE | 31 January 2021, 22:32:18 UTC |
5cbd6e9 | Alexander Kruppa | 31 January 2021, 16:31:18 UTC | Avoid warnings with -Wtype-limits and ASSERT_EXPENSIVE | 31 January 2021, 22:32:15 UTC |
b60d527 | Emmanuel Thomé | 31 January 2021, 21:45:47 UTC | Merge branch 'various-fixes' into 'master' Various fixes Closes #30010 See merge request cado-nfs/cado-nfs!20 | 31 January 2021, 21:45:47 UTC |
5fc870d | Emmanuel Thomé | 31 January 2021, 12:44:58 UTC | fix race condition on chmod in cado-nfs-client | 31 January 2021, 21:33:01 UTC |
4da96c5 | Emmanuel Thomé | 31 January 2021, 01:18:41 UTC | fix some c++ pedantry (failure on arm32 abi) | 31 January 2021, 21:32:45 UTC |
6f588f0 | Emmanuel Thomé | 31 January 2021, 01:18:37 UTC | minor | 31 January 2021, 21:32:40 UTC |
a843800 | Emmanuel Thomé | 30 January 2021, 23:03:06 UTC | detect musl libc, fix oddity in sqrt ; fix #30010 | 31 January 2021, 21:32:14 UTC |
195af8c | Emmanuel Thomé | 30 January 2021, 14:40:21 UTC | report dead client output on client failure (for spawned clients) | 31 January 2021, 21:31:36 UTC |
ec8dbdb | Emmanuel Thomé | 30 January 2021, 16:40:25 UTC | polyselect output must always be present if it gets parsed | 31 January 2021, 21:31:31 UTC |
a1d80df | Emmanuel Thomé | 31 January 2021, 19:59:56 UTC | Merge branch '30011-cado-nfs-client-py-dies-with-valueerror-check_hostname-needs-a-ssl-context-with-either' into 'master' Resolve "cado-nfs-client.py dies with "ValueError: check_hostname needs a SSL context with either CERT_OPTIONAL or CERT_REQUIRED" with Python 3.6" Closes #30011 See merge request cado-nfs/cado-nfs!19 | 31 January 2021, 19:59:56 UTC |
c086c66 | Alexander Kruppa | 31 January 2021, 19:03:56 UTC | Set ssl.Context.check_hostname after verify_mode to avoid ValueError on Python < 3.7. Fixes #30011 | 31 January 2021, 19:05:18 UTC |
f71bf45 | Paul Zimmermann | 31 January 2021, 05:48:17 UTC | hopefully this version is now correct | 31 January 2021, 05:48:17 UTC |
80b7d53 | Paul Zimmermann | 30 January 2021, 17:31:37 UTC | definitely fix commit cb807b0 which was nonsense | 30 January 2021, 17:31:37 UTC |
2d4274b | Emmanuel Thomé | 30 January 2021, 12:42:25 UTC | add a call interface to filter_rels in order to get sorted indices relations out of earlyparser_need_index_sorted are sorted if they're not sorted already | 30 January 2021, 14:10:52 UTC |
68edfa5 | Emmanuel Thomé | 30 January 2021, 11:07:53 UTC | this rbfl cannot use a reference | 30 January 2021, 11:56:59 UTC |
c4b8033 | Emmanuel Thomé | 30 January 2021, 11:01:23 UTC | refresh gf2x -- some important fixes are in the main repo | 30 January 2021, 11:56:53 UTC |
7348d48 | Emmanuel Thomé | 28 January 2021, 21:48:48 UTC | PAGE_SIZE is sometimes defined by libc (alpine linux) | 29 January 2021, 15:41:07 UTC |
cb807b0 | Paul Zimmermann | 28 January 2021, 08:34:05 UTC | [filter/merge.c] check the input rows are sorted (this can be removed once that check is done in filter_rels) | 28 January 2021, 08:34:05 UTC |
98aef21 | Paul Zimmermann | 28 January 2021, 07:30:03 UTC | [merge] added a check when no merge is done while n_possible_merges > 0 (this should not happen if the input purged file is valid) | 28 January 2021, 07:30:03 UTC |
d14840c | Emmanuel Thomé | 27 January 2021, 21:13:04 UTC | Merge branch 'various-fixes' into 'master' Various fixes See merge request cado-nfs/cado-nfs!16 | 27 January 2021, 21:13:04 UTC |
424a304 | Emmanuel Thomé | 26 January 2021, 15:21:51 UTC | fix c++ bug revealed by clang11 | 27 January 2021, 20:59:11 UTC |
2b71ce3 | Emmanuel Thomé | 26 January 2021, 14:28:45 UTC | ClientServerTask ; die if localhost clients die. This effectively aborts the computation if a localhost client dies. Overall, this idea of clients spawned by the server is deprecated in my opinion, so it's not a big deal. Most often, when such a client dies, it's something that gets in our way within tests (e.g. a test that never returns because it waits for the timeout to expire, which is due to happen if the real even is that the client has just died). I think it's better that way (behaviour restricted to localhost client, remember). But it *might* be disruptive in certain cases. | 27 January 2021, 20:59:07 UTC |
331f886 | Emmanuel Thomé | 26 January 2021, 14:27:57 UTC | cado-nfs-client fix: catch all output in daemon mode It is still slightly annoying that the client failure is not always reported to the server, but at least if the python code errors out, at least the backtrace is *somewhere* | 27 January 2021, 20:59:01 UTC |
b9d4069 | Emmanuel Thomé | 25 January 2021, 22:20:33 UTC | fix overzealous have_mmx=1 on 64-bit | 27 January 2021, 20:56:26 UTC |
de42778 | Pierrick Gaudry | 13 December 2020, 08:35:48 UTC | Replace python2 Thread.isAlive() by python3 is_alive() | 13 December 2020, 08:35:48 UTC |
8ef06dd | Pierrick Gaudry | 09 December 2020, 20:56:47 UTC | Update Readme.Python according to current requirements. | 09 December 2020, 20:56:47 UTC |
542275c | Pierrick Gaudry | 09 December 2020, 20:34:33 UTC | fractions.import is now deprecated. | 09 December 2020, 20:34:33 UTC |
be718f1 | Emmanuel Thomé | 02 December 2020, 17:07:58 UTC | Merge branch 'use-relation-cache' into 'master' Use relation cache See merge request cado-nfs/cado-nfs!15 | 02 December 2020, 17:07:58 UTC |
615d0ad | Emmanuel Thomé | 02 December 2020, 16:11:22 UTC | fix printing bug | 02 December 2020, 16:11:22 UTC |
30597bd | Emmanuel Thomé | 02 December 2020, 16:11:10 UTC | set a default value for maxbits in estimate_matsize | 02 December 2020, 16:11:10 UTC |
e13a9b1 | Emmanuel Thomé | 02 December 2020, 13:57:45 UTC | change the relation cache search logic so as to support cross-boundary ranges | 02 December 2020, 13:57:45 UTC |
e262818 | Emmanuel Thomé | 02 December 2020, 10:30:56 UTC | add relation-cache option to random-sample mode in las Tehe relation cache can be created with the script: tests/estimate_matsize/build_relation_cache.pl For large runs where "real" relations are available, this accelerates the creation of sample relations significantly. | 02 December 2020, 13:06:46 UTC |
4385423 | Emmanuel Thomé | 02 December 2020, 10:00:09 UTC | Merge branch 'add-simplistic-json-parser' into use-relation-cache | 02 December 2020, 10:00:09 UTC |
78a2bd8 | Emmanuel Thomé | 01 December 2020, 19:29:45 UTC | q&d json parser and printer | 02 December 2020, 09:57:44 UTC |
72658b9 | Emmanuel Thomé | 28 November 2020, 14:16:53 UTC | add a test for estimate_matsize + comp sq | 28 November 2020, 14:16:53 UTC |
f15ee50 | Emmanuel Thomé | 27 November 2020, 22:13:41 UTC | Merge branch 'nocncheck-python3' into 'master' implement --nocncheck for python3 cado-nfs-client See merge request cado-nfs/cado-nfs!14 | 27 November 2020, 22:13:41 UTC |
0cb1e01 | Emmanuel Thomé | 27 November 2020, 22:00:07 UTC | implement --nocncheck for python3 cado-nfs-client | 27 November 2020, 22:00:07 UTC |
6605b6e | Emmanuel Thomé | 27 November 2020, 21:35:19 UTC | Merge branch 'use-temp-name-when-writing' into 'master' Use temporary file names when creating output. See merge request cado-nfs/cado-nfs!13 | 27 November 2020, 21:35:19 UTC |
d659280 | Emmanuel Thomé | 27 November 2020, 21:18:54 UTC | Use temporary file names when creating output. This changes avoides the pitfall of partially written output file that are indistinguishable from the expected final files. This only works when the C++ ofstream_maybe_compressed layer is used. (doing the same with the C functions would require a more significant interface change). An illustration can be: ``` strace -f -e trace=open,openat,rename ./build/localhost/sieve/freerel -poly ./tests/misc/test_renumber.data/mnfs5.poly -renumber /tmp/renumber.gz -out /dev/null -lpbs 11,10,10,10,10 -lcideals > /dev/null [pid 124055] openat(AT_FDCWD, "/tmp/renumber.gz.tmp.124054", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 1 [...] [pid 124054] rename("/tmp/renumber.gz.tmp.124054", "/tmp/renumber.gz") = 0 ``` | 27 November 2020, 21:18:54 UTC |
3dc7496 | Pierrick Gaudry | 25 November 2020, 14:12:43 UTC | Too noisy sm_simple. | 25 November 2020, 14:12:43 UTC |
3a63c87 | Pierrick Gaudry | 01 November 2020, 13:39:51 UTC | Make the option -no-recomp-norm of finishbatch more robust. | 01 November 2020, 13:39:51 UTC |
7fe0fdc | Pierrick Gaudry | 28 October 2020, 07:32:49 UTC | Fix previous commit | 28 October 2020, 07:32:49 UTC |
3e0a4d2 | Pierrick Gaudry | 27 October 2020, 15:55:37 UTC | Add an obscure option to finishbatch, in case full norms are passed. | 27 October 2020, 15:55:37 UTC |
8ab2eea | Emmanuel Thomé | 21 October 2020, 08:29:52 UTC | Merge branch 'fix-30009' into 'master' scale algebraic square by $`{\hat f}'(\hat\alpha)^2`$ See merge request cado-nfs/cado-nfs!12 | 21 October 2020, 08:29:52 UTC |
7c10b68 | Emmanuel Thomé | 21 October 2020, 07:56:30 UTC | scale algebraic square by $`{\hat f}'(\hat\alpha)^2`$ | 21 October 2020, 07:57:21 UTC |
cfc03e7 | Emmanuel Thomé | 20 October 2020, 10:41:13 UTC | Merge branch 'mpz-poly-parallel-onlyifneeded' into master | 20 October 2020, 10:41:13 UTC |
bc21078 | Emmanuel Thomé | 20 October 2020, 09:30:19 UTC | implement transparent interface for mpz_poly parallel code | 20 October 2020, 10:22:43 UTC |
9b86829 | Emmanuel Thomé | 19 October 2020, 17:00:13 UTC | mark as MAYBE_UNUSED for cases where we have no openmp | 20 October 2020, 07:30:22 UTC |
82b9bc1 | Emmanuel Thomé | 19 October 2020, 16:59:34 UTC | style fix | 20 October 2020, 07:30:11 UTC |
d29140f | Emmanuel Thomé | 19 October 2020, 16:44:11 UTC | do not disable the mpz_poly parallel code in sqrt accumulation! | 20 October 2020, 07:15:09 UTC |
e12f34d | Emmanuel Thomé | 19 October 2020, 21:18:11 UTC | there's no zcat on osx. use gzip -dc instead | 19 October 2020, 21:19:24 UTC |
297aa71 | Emmanuel Thomé | 19 October 2020, 16:49:17 UTC | Merge branch 'make-estimate-matsize-idempotent' into 'master' Fixes to estimate_matsize.sh See merge request cado-nfs/cado-nfs!10 | 19 October 2020, 16:49:17 UTC |
30f520e | Emmanuel Thomé | 19 October 2020, 16:25:00 UTC | We use bc in some scripts | 19 October 2020, 16:25:00 UTC |
d0dd50e | Emmanuel Thomé | 19 October 2020, 08:14:35 UTC | expose mpz_poly interface in two different ways. | 19 October 2020, 16:19:58 UTC |
dbb9c76 | Emmanuel Thomé | 19 October 2020, 15:16:43 UTC | tests for estimate_matsize | 19 October 2020, 15:23:19 UTC |
bddb4c3 | Emmanuel Thomé | 19 October 2020, 15:15:24 UTC | corner case in estimate_matsize | 19 October 2020, 15:22:03 UTC |
83b8b30 | Emmanuel Thomé | 19 October 2020, 15:21:55 UTC | Merge branch 'master' into make-estimate-matsize-idempotent | 19 October 2020, 15:21:55 UTC |
24e3f8e | Emmanuel Thomé | 19 October 2020, 15:16:29 UTC | protection for empty ranges in las --random-sample | 19 October 2020, 15:21:41 UTC |
3d1f7a1 | Emmanuel Thomé | 19 October 2020, 15:15:56 UTC | corner cases in las (for *really* small special-q) | 19 October 2020, 15:21:37 UTC |
1b06e32 | Emmanuel Thomé | 19 October 2020, 15:15:10 UTC | fix bug in merge | 19 October 2020, 15:19:02 UTC |
d9072a7 | Emmanuel Thomé | 19 October 2020, 14:05:45 UTC | Fixes to estimate_matsize.sh | 19 October 2020, 14:05:45 UTC |
b4a8ead | Emmanuel Thomé | 19 October 2020, 13:24:47 UTC | Merge branch 'make-estimate-matsize-idempotent' into 'master' Make estimate matsize idempotent See merge request cado-nfs/cado-nfs!9 | 19 October 2020, 13:24:47 UTC |
81489bf | Emmanuel Thomé | 19 October 2020, 13:10:48 UTC | Arrange so that estimate-matsize reuses already computed files This is done in a pretty simplistic way. If the number of threads changes, the files are computed again. | 19 October 2020, 13:10:48 UTC |
29bcfab | Paul Zimmermann | 16 October 2020, 14:21:01 UTC | Merge branch 'master' of gitlab.inria.fr:cado-nfs/cado-nfs into master | 16 October 2020, 14:21:01 UTC |
196f4d6 | Paul Zimmermann | 16 October 2020, 14:20:47 UTC | fixed typo | 16 October 2020, 14:20:47 UTC |
17b87f4 | Emmanuel Thomé | 16 October 2020, 12:20:25 UTC | add ref | 16 October 2020, 12:20:25 UTC |
9a2d0bf | Paul Zimmermann | 15 October 2020, 16:21:04 UTC | added some debug code | 15 October 2020, 16:21:04 UTC |
b8b1c9f | Paul Zimmermann | 15 October 2020, 15:13:35 UTC | make -fsanitize=undefined -fno-sanitize-recover=all happy (although it looks like a false report) | 15 October 2020, 15:13:35 UTC |
24225aa | Paul Zimmermann | 15 October 2020, 15:13:18 UTC | Merge branch 'master' of gitlab.inria.fr:cado-nfs/cado-nfs | 15 October 2020, 15:13:18 UTC |
ad274c0 | Emmanuel Thomé | 15 October 2020, 11:54:22 UTC | compilation fixes for old g++ | 15 October 2020, 11:54:39 UTC |
db93fba | Paul Zimmermann | 15 October 2020, 11:22:16 UTC | Merge branch 'master' of gitlab.inria.fr:cado-nfs/cado-nfs | 15 October 2020, 11:22:16 UTC |
33c7ce9 | Paul Zimmermann | 15 October 2020, 11:21:46 UTC | use "unsigned long" for valuation in scaled polynomials | 15 October 2020, 11:21:46 UTC |
98f6fb9 | Paul Zimmermann | 15 October 2020, 11:20:38 UTC | forgot to double k in previous commit | 15 October 2020, 11:20:38 UTC |
70dce17 | Paul Zimmermann | 15 October 2020, 10:32:25 UTC | misplaced invariant check | 15 October 2020, 10:32:25 UTC |
6936eca | Paul Zimmermann | 15 October 2020, 08:01:58 UTC | Merge branch 'master' of gitlab.inria.fr:cado-nfs/cado-nfs | 15 October 2020, 08:01:58 UTC |
c1edab4 | Paul Zimmermann | 15 October 2020, 08:01:30 UTC | make -nratchars work | 15 October 2020, 08:01:30 UTC |
6a75e69 | Emmanuel Thomé | 14 October 2020, 13:35:52 UTC | Allow usage of estimate_matsize.sh with -params arg as in: ./scripts/estimate_matsize.sh -params /tmp/params.txt foo.poly | 14 October 2020, 13:36:41 UTC |
9524ebc | Paul Zimmermann | 12 October 2020, 14:10:59 UTC | added link between releases and commit numbers | 12 October 2020, 14:10:59 UTC |
f8e1f33 | Paul Zimmermann | 07 October 2020, 11:37:15 UTC | make it clear which version of KB/MB we consider (cf https://en.wikipedia.org/wiki/Mebibyte) | 07 October 2020, 11:37:15 UTC |
c82308d | Paul Zimmermann | 05 October 2020, 13:15:49 UTC | populate .mailmap | 05 October 2020, 13:15:49 UTC |
8d0561c | Paul Zimmermann | 28 September 2020, 11:32:04 UTC | typo | 28 September 2020, 11:32:04 UTC |
0231106 | Emmanuel Thomé | 28 September 2020, 10:57:13 UTC | test | 28 September 2020, 10:57:13 UTC |
8c681c4 | Pierrick Gaudry | 28 September 2020, 07:48:53 UTC | more doc for tests | 28 September 2020, 07:50:02 UTC |
51f3d52 | Pierrick Gaudry | 28 September 2020, 07:48:31 UTC | Another test for fake_rels | 28 September 2020, 07:50:02 UTC |
1bee210 | Paul Zimmermann | 25 September 2020, 08:11:19 UTC | removed useless cast | 25 September 2020, 08:11:19 UTC |
9d52c08 | Pierrick Gaudry | 25 September 2020, 07:15:46 UTC | signedness mayhem in fake_rels | 25 September 2020, 07:15:46 UTC |
445a9f6 | Emmanuel Thomé | 23 September 2020, 06:26:31 UTC | Merge branch 'master' of gitlab.inria.fr:cado-nfs/cado-nfs into master | 23 September 2020, 06:26:31 UTC |
27ef73b | Emmanuel Thomé | 23 September 2020, 06:26:26 UTC | clarification in comment | 23 September 2020, 06:26:26 UTC |
e32c0c4 | Paul Zimmermann | 18 September 2020, 08:03:56 UTC | [dup2] hard-code the number of threads (for roots mod p) to 3 | 18 September 2020, 08:03:56 UTC |
08e5842 | Emmanuel Thomé | 17 September 2020, 09:51:52 UTC | Merge branch 'deprecate-unversioned-python' into 'master' remove unversioned #! python in cado-nfs-client.py See merge request cado-nfs/cado-nfs!6 | 17 September 2020, 09:51:52 UTC |
21dc5e8 | Emmanuel Thomé | 17 September 2020, 09:36:39 UTC | use flag created by 2d3879e5b in the gitlab pipeline | 17 September 2020, 09:36:39 UTC |
ac7df46 | Emmanuel Thomé | 17 September 2020, 09:27:23 UTC | fix weird logic in cado-nfs.py's main | 17 September 2020, 09:27:23 UTC |
cfd499d | Emmanuel Thomé | 17 September 2020, 08:44:05 UTC | remove unversioned #! python in cado-nfs-client.py This addresses #30007, and effectively removes support for python2 clients. | 17 September 2020, 08:44:05 UTC |
179bd32 | Emmanuel Thomé | 17 September 2020, 05:30:11 UTC | Merge branch 'fix-weird-timing-reports' into 'master' Fix weird timing reports https://lists.gforge.inria.fr/pipermail/cado-nfs-discuss/2020-September/001302.html Several inconsistent timing reports. This fixes some of the inconsistencies mentioned in #30008 ; perhaps not all See merge request cado-nfs/cado-nfs!5 | 17 September 2020, 05:30:11 UTC |
0cd9d87 | Emmanuel Thomé | 16 September 2020, 22:07:48 UTC | report aggregated CPU time for krylov and mksol as well. | 16 September 2020, 22:07:48 UTC |
3e667af | Emmanuel Thomé | 16 September 2020, 21:50:38 UTC | don't try to report stats for tasks that don't run | 16 September 2020, 21:50:38 UTC |
2d3879e | Emmanuel Thomé | 16 September 2020, 21:11:50 UTC | fix a timing parse error, + add detection flag. A slight change in GF(2) linear algebra lingen output was causing the linear algebra time to be misreported for factoring. | 16 September 2020, 21:11:50 UTC |
1eec3e5 | Emmanuel Thomé | 16 September 2020, 20:48:53 UTC | Fix minor bug with progress reporting in polyselect tasks progress estimates based on wu_received were not robust to potential modificaions of adrange in the course of the computation. | 16 September 2020, 20:48:53 UTC |
ac03d61 | Pierrick Gaudry | 08 September 2020, 08:16:36 UTC | Fix bug that forced recomputation of SMs | 08 September 2020, 08:16:36 UTC |