https://github.com/linbox-team/fflas-ffpack

sort by:
Revision Author Date Message Commit Date
2b1367f fix round in test-simd.C + more explicit rounding mode in Simd512 structs 24 November 2020, 18:09:56 UTC
563c3e0 test-simd.C: remove unused macro 24 November 2020, 15:28:49 UTC
3de7a17 Simd: add blendv test + clean up blendv method in simd512 structs 24 November 2020, 11:24:20 UTC
b4a8eca test-simd.C: do not consider const references as inputs for tests 24 November 2020, 11:03:56 UTC
627da0c test-simd.C: fix width for pretty printing debug data of tests 24 November 2020, 11:02:30 UTC
de0abc9 test-simd.C: change random distribution for floating point tests 24 November 2020, 11:01:43 UTC
6a8bd05 test-simd.C : last rewrite 19 November 2020, 09:12:59 UTC
1d48cd3 test-simd.C : small rewrite to ease the reuse with future SimdModular 18 November 2020, 11:41:46 UTC
500dff8 Simd: fix blend in some Simd struct + test blend in test-simd.C + tidy up comments of blend methods 17 November 2020, 16:08:16 UTC
deb1ab5 test-simd: forgot to remove one piece of comment 17 November 2020, 15:50:20 UTC
e6c6e85 Improve test-simd 16 November 2020, 14:20:28 UTC
8f3ecda simd: fix fmsubxin for Simd128 on 64-bit integers (const and & missing) 16 November 2020, 12:56:36 UTC
c829b95 Simd: add unpacklohi and pack methods (#321) * simd: add unpacklohi method in simd struct for simd128 and simd256 * simd: more consistent interface for unpacklo/hi, pack, ... (WIP) + add tests * simd256_int16.inl: remove comments that should have not be there * simd: remove unpacklo128 and unpackhi128 from simd256 struct * simd: add unpacklo, unpackhi, pack_even, pack_odd methos in simd521 structs * simd512: replace loadu_epi32 by loadu_si512 * simd512 for float: rewrite pack_even, pack_odd and pack * simd: fix lots of typos 05 November 2020, 17:03:51 UTC
539d47b Add more precise type_string to SIMD structs (#320) 05 November 2020, 16:44:03 UTC
fd9f042 Fix for greater and lesser in simd (+test in test-simd.C) #317 (#318) 05 November 2020, 15:48:39 UTC
0147210 call SET_FLAGS from configure.ac #315 (#316) 28 August 2020, 09:18:59 UTC
3cd3145 gfops irrelevant to please perfpublisher 26 August 2020, 09:10:49 UTC
2e653f7 no CXXFLAGS for C compiler 26 August 2020, 07:58:09 UTC
1da2083 make output suite CI requirements 26 August 2020, 07:47:32 UTC
1868602 fix missing inline 26 August 2020, 07:39:23 UTC
a677c35 Bruhat gen (#311) * first draft for bruhat gen * add test file * start writing a random generator of Left Triangular QuasiSep matrices * Update of Bruhat Generator * Update of Bruhat Generator 2 * Update of Bruhat Generator 3 * random generation of LTQS RPM * Update of Bruhat Generator * fzero for triangular matrices + left and right triangular tag * fix call to fzero * Update of Bruhat Generator * header in ffpack * typo * fix missing test * random gen now considers the QS order param * randgen now works * compilation fixes * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * remove endif * Update of Bruhat Generator * Update of Bruhat Generator * fix compilation errors * Update of Bruhat Generator * done with the rand generation (no longer any infinite loop * Update of Bruhat Generator * Computing the permutation to echelon form * fix * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * compress permutation is a MathPerm * fix leak * M allocated by callee * fix perm size * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of Bruhat Generator * Update of BruhatxTS Matrix * Update of Bruhat product * Update of Bruhat product * Update of Bruhat product * Update of Bruhat product * Update of Bruhat product * doc on permutations * Update of Bruhat product * Update of Bruhat product * Update of Bruhat product * fix l->m dimension * debug in progress * fix bug with rs * fix bug * fixed a bug * fixed ! * new benchmark for quasisep matrices * add bruhatgen file to install * doc + indent + order of args * cleanup * cleanup in progress * cleanup ++ * done cleanup of partial sums * done reviewing * bib ref Co-authored-by: Clément Pernet <clement.pernet@gmail.com> 31 July 2020, 13:52:40 UTC
e152f01 Merge branch 'master' of github.com:linbox-team/fflas-ffpack 11 June 2020, 12:11:35 UTC
11695cf fscal 10 June 2020, 08:00:26 UTC
23928f9 remove 1st and last file in .pc file (#308) this causes problem if building with ksh, as they remain, causing a broken .pc file. 09 May 2020, 11:48:59 UTC
f1f9ee7 fabi-version=6 only if gcc version < 5 (#305) * fabi-version=6 only if gcc version < 5 * fabi-version=6: detect gcc >= 5, gcc 4.3+ and gcc 4.9.2 Co-authored-by: Cyril Bouvier <cyril.bouvier@lirmm.fr> 17 March 2020, 12:09:04 UTC
cad4a57 autotools: towards uniformization of givaro and fflas-ffpack (#293) * autotools: First step towards uniformization of givaro and fflas * autotools: fix for gcc 4.9 * fabi-version=6: detect gcc >= 5, gcc 4.3+ and gcc 4.9.2 Co-authored-by: Jean-Guillaume Dumas <Jean-Guillaume.Dumas@imag.fr> 17 March 2020, 11:41:09 UTC
2d9df94 fix bug for RNS conversion with modular<ruint> (#306) Solves issue #300. 16 March 2020, 14:59:33 UTC
d73feae symmetric freduce only on square matrices, and actually use it in fsyrk_strassen (#303) 13 February 2020, 09:31:11 UTC
cc6adc2 Charpoly checker over the integers is not yet implemented (#302) 13 February 2020, 08:37:01 UTC
c99b22f Delayed pre-computations in fsyrk-Strassen (#299) * WIP: new fsyrk_strassen * schedule for fsyk_strassen with accumulation and 3 temps * commit at 10000m * computeS1S2 * fix comment * compiling * still buggy * schedule with no acc working on dimension power of 2 * schedule with accumulation now works! (with k=n=2^m). * no verbosity * done machinery to reduce to cblas_xsyrk * done * ../fflas-ffpack/fflas/fflas_fsyrk_strassen.inl * fix wrong trans and gfops * clearer log * remove verbosity * add syrk_strassen to benchmark * improved computeS1S2 * no randiter when no field * done fixing bugs * slicing for odd dimensions * fix comment * fix bug in AATpC and change sign of S2 * new schedule + WIP on delayed * done with delayed * delayed computeS1S2 * fixed all bugs * remove verbosity * done * fix bug * Symmetric freduce * fix and benchmark classic divide and conquer fsyrk * new algo trait D&C * fix update of lazy helper * fsyrk Upper and Trans: 1st step * fix calls in test suite * fsyrk upper wip * making progress * fixed for Upper and Lower NoTrans * first step in fsyrk-trans, still not working * trans WIP * fix k=0 check * working back again * include file in distrib * fixes to trans * done debugging trans * fix k>n bug * fix default case * no int32_t * no helper * clean up benchmark * cleanup * cleanup * minor fixes * add a test for computeS1S2 * test computeS1S2 trans * control of the delayed mod reductions in computeS1S2 * NeedPreXXXReduction * WIP fixing bugs * fixed for modular<float> * trans wors * almost full test suite * check bounds on triangular matrices * fixed last (?) bugs * fixed last remaining bugs * done debuggind: the modular<int64_t>::maxcard issue is too deep into the whole library to be fixed here: disabling * cleanup debugging verbosity * fix maxcard * remove forceField * freduce UpLo works only if m<=n for now Co-authored-by: Jean-Guillaume Dumas <jgdumas@users.noreply.github.com> 12 February 2020, 17:08:54 UTC
07f3c7f Update MaxCardinality to MaxFFLASCardinality (#290) * rewrite of ChooseField using the new maxFFLASCardinality in Givaro * explicit cast to please clang ambiguous call to either cast or constructor * disable igemm with kmax < 4 * transpose * inline to fix multilib compilation failure * remove verb * double * parenthesis * print seed only when failed 12 February 2020, 15:11:08 UTC
34df86e fix bug 31 January 2020, 14:32:07 UTC
9c2e5c4 Threadindex (#281) * Added firstprivate for iterator to TASK in FOR1D * VALUE instead of firstprivate * new THREAD_INDEX construct 17 December 2019, 08:40:26 UTC
aa5e397 Added firstprivate for iterator to TASK in FOR1D (#280) * Added firstprivate for iterator to TASK in FOR1D * VALUE instead of firstprivate 17 December 2019, 08:39:41 UTC
2d3af1a BLAS based base-case for fsyrk (#283) * done machinery to reduce to cblas_xsyrk * done * fix wrong trans and gfops 17 December 2019, 08:35:27 UTC
1d2e2b7 updating GIVARO requirements to 4.1.1 (#287) 16 December 2019, 15:16:31 UTC
02f63b9 Adding public attributes to Simd structs (#269) * simd: add extra public "using" to ease the usage of aligned vector * simd: add extra public "using" to ease checking compatibility with Modular 10 September 2019, 14:05:41 UTC
86cd45f fix missing return 09 September 2019, 13:36:18 UTC
e405785 _internal_iterator cannot be modified within the TASK call. (#278) call begin(), end() beforehand. No those are sequential, even if TASKS remai parallel 28 August 2019, 14:22:19 UTC
91dc2f3 Update README.md 20 August 2019, 09:02:01 UTC
7b122a3 Add the option of MODE for the use of Paladin FOR1D (#275) * Add the option of MODE for the use of Paladin FOR1D * Removed stdio in the parallel.h * Deleted commented FOR1D without MODE * Cleaned up according to the required changes and ready for code review 05 August 2019, 09:09:30 UTC
244fec7 Plug memory leaks. (#276) 23 July 2019, 08:47:17 UTC
2ac7c70 bitsize --> uint64_t 10 July 2019, 13:36:10 UTC
2b5b5f3 size_t --> uint64_t 10 July 2019, 12:43:35 UTC
69c6b04 fix travis 04 July 2019, 09:08:36 UTC
06f82f5 charpoly autotuning with larger field, to avoid Las-Vegas failure 01 July 2019, 09:17:39 UTC
118d7ef setSeed is now FFLAS::setSeed() (#271) samplesize from Integer 24 June 2019, 09:23:30 UTC
c309966 seed/size order reverted (#272) * seed/size order reverted * ok no bitSize nor samplesize for now on the subfields of a RNS * also no bitsize 24 June 2019, 09:22:41 UTC
e723e6a make autotune should fail when some run segfaults so Jenkins can detect it (#270) 21 June 2019, 13:45:24 UTC
c2db94f Randiter refacto (#263) * start randiter refacto (samplesize and order seed first) * bits are used to choose field, defautlt randiter will use them * spacing * noexec 21 June 2019, 08:21:49 UTC
9a9aa42 absCompare saves a large factor w.r.t. bitsize (#268) * absCompare saves a large factor w.r.t. bitsize * givaro * oups lda 12 June 2019, 15:03:52 UTC
01ced8e version++ 11 June 2019, 11:21:37 UTC
469a041 fix bug arithprog 11 June 2019, 11:18:48 UTC
b7e4bc2 fix p*RankProfile API 11 June 2019, 08:25:49 UTC
0b6d2b5 preparin new minor relelease 11 June 2019, 08:07:56 UTC
d54106c change p*Echelonform API 11 June 2019, 08:01:23 UTC
2731e92 passing numthreads to all p.. functions 07 June 2019, 14:24:16 UTC
2ac1a02 preparing release 2.4.1 07 June 2019, 13:22:39 UTC
079b4c7 Update README.md 07 June 2019, 08:09:57 UTC
69ad8cb add README.md to distro 07 June 2019, 08:07:48 UTC
ddba4cc add missing set_numthreads 29 May 2019, 07:59:00 UTC
c4c39f6 version++ 27 May 2019, 11:37:15 UTC
157522b revert to linbox-osx 20 May 2019, 15:34:00 UTC
289a3c3 Merge pull request #264 from hygonsoc/master queryCacheSizes: add Hygon Dhyana support to fix cache size bug 20 May 2019, 11:01:23 UTC
5cc2c14 queryCacheSizes: add Hygon Dhyana support to fix cache size bug Signed-off-by: fanjinke <fanjinke@hygon.cn> 20 May 2019, 04:15:07 UTC
b7c14b2 update to new CI OSX slave 15 May 2019, 12:43:18 UTC
f064ed6 set xrows ldx when reading the B matrix 10 May 2019, 15:26:53 UTC
18e02fe Merge pull request #262 from linbox-team/fewer_openblas_setnumthreads enable a prologue setting openblas numthreads once for all 10 May 2019, 14:49:35 UTC
2542c9b Update README.md 10 May 2019, 14:36:25 UTC
b5073b8 enable a prologue setting openblas numthreads once for all 10 May 2019, 13:46:46 UTC
93eae1e reducing the load of the test suite 10 May 2019, 12:46:08 UTC
51a9bd1 ready for review 10 May 2019, 09:09:11 UTC
895e244 merge CONTRIBUTORS in AUTHORS 10 May 2019, 09:05:44 UTC
a625edc Merge pull request #261 from linbox-team/no_simd_for_cygwingcc Disable AVX with gcc on CYGWIN. 10 May 2019, 08:07:59 UTC
88df911 Update fflas-ffpack-config.h 09 May 2019, 20:25:10 UTC
e16f4ef prevent using >128bits registers with gcc on CYGWIN 09 May 2019, 19:07:48 UTC
a3362b6 update charpoly autotuner 03 May 2019, 12:52:55 UTC
249644a update precompiled libs 03 May 2019, 12:38:35 UTC
2ced4e0 rc1 03 May 2019, 12:23:32 UTC
29f87d1 update changelog 03 May 2019, 12:23:09 UTC
db49647 Merge pull request #260 from linbox-team/artihprog-separate-precond ArithProg Charpoly: separate preconditionner from main algorithm 03 May 2019, 12:17:28 UTC
bce63aa Merge pull request #256 from linbox-team/parallel-RankProfile-encapsulation-of-ParSeqHelper Wrapped Rank as well as bother Row and Column Rankprofile with no longer the need of using a PAR_BLOCK to label the parallel region 03 May 2019, 12:14:15 UTC
59ff98d Merge branch 'parallel-RankProfile-encapsulation-of-ParSeqHelper' of https://github.com/linbox-team/fflas-ffpack into parallel-RankProfile-encapsulation-of-ParSeqHelper 03 May 2019, 07:55:40 UTC
83fd234 Cleaned up the test-rankprofile according to comments after code review 03 May 2019, 07:54:52 UTC
5ebce5e Merge pull request #251 from linbox-team/fgemv-with-ParSeqHelper Fgemv with par seq helper 02 May 2019, 16:09:09 UTC
f603cf0 Merge pull request #252 from linbox-team/parallel-solve-encapsulation-of-ParSeqHelper self-standing parallel routine named pSolve for the routine Solve added and cleaned up 02 May 2019, 16:04:21 UTC
f3de754 Merge pull request #254 from linbox-team/parallel-det-encapsulation-of-ParSeqHelper Wrapped Det into pDet so that no need to use PAR_BLOCK to label the parallel region 02 May 2019, 16:02:58 UTC
e198d81 Merge pull request #255 from linbox-team/parallel-pluq-encapsulation-of-ParSeqHelper Wrapped PLUQ into pPLUQ without the need to use PAR_BLOCK for labelling the parallel region 02 May 2019, 16:01:17 UTC
4341e2e tilrecursive by default for parallel rank profiles 02 May 2019, 15:57:24 UTC
6213fc2 Merge pull request #257 from linbox-team/parallel-EchelonForm-encapsulation-of-ParSeqHelper Wrappped EchelonForm related functions into their parallel routines with no more need for PAR_BLOCK 02 May 2019, 15:49:51 UTC
e7ffbc3 add precompiled interfaces for parallel echelon 02 May 2019, 15:16:23 UTC
d4fa581 tile recursive by default for parallel echelon 02 May 2019, 15:16:03 UTC
f53c3df Wrappped EchelonForm related functions into their parallel routines with no more need for PAR_BLOCK 02 May 2019, 12:03:10 UTC
4e89e05 Wrapped Rank as well as bother Row and Column Rankprofile with no longer the need of using a PAR_BLOCK to label the parallel region 02 May 2019, 10:56:51 UTC
96cefb7 Wrapped PLUQ into pPLUQ without the need to use PAR_BLOCK for labelling the parallel region 02 May 2019, 09:21:13 UTC
6a73e1d Wrapped Det into pDet so that no need to use PAR_BLOCK to label the parallel region 02 May 2019, 07:48:44 UTC
ae22bf1 cleaned up the test-solve according to comments after code review 30 April 2019, 15:53:56 UTC
fa644c2 Restrcutured the fgemv for sequential helper according to the comments after code review 30 April 2019, 15:42:13 UTC
8337c27 cleaned up the test-solve 30 April 2019, 13:37:05 UTC
7dcd386 Cleaned up according to comments after code review 30 April 2019, 13:07:12 UTC
back to top