| e6b9677 | ZHG | 15 November 2017, 12:02:14 UTC | Multiple data type test for hybride parallelization of fgemv with <Recursive,Thread> and <Row,Thread> MMHelper added but to avoid processes shared among cores better to use -bind-to-core with mpirun command especially for 2 processes | 15 November 2017, 12:02:14 UTC |
| 0afc262 | ZHG | 14 November 2017, 15:45:04 UTC | No important modification | 14 November 2017, 15:45:04 UTC |
| 3785c82 | ZHG | 14 November 2017, 15:43:06 UTC | Test with row cutting strategy for the hybride pfgemv added but need to be clean up later | 14 November 2017, 15:43:06 UTC |
| ee433f7 | ZHG | 13 November 2017, 17:03:17 UTC | Correct matrix splitting bugs for the shared memory computing openMP block | 13 November 2017, 17:03:17 UTC |
| 5834d44 | ZHG | 13 November 2017, 16:54:57 UTC | Rewrite the test program for the hybride parallelization of fgemv as the previous one has been lost PS: still need to work on the test on different pfgemv for the hybride parallelization test | 13 November 2017, 16:54:57 UTC |
| dcfaa6d | ZHG | 13 November 2017, 12:28:45 UTC | Modified Makefile.am to compile with mpicc/mpic++ | 13 November 2017, 12:28:45 UTC |
| ddf1fd9 | ZHG | 13 November 2017, 10:54:50 UTC | Improved hybride parallelization with openMP and MPI for fgemv added: no more infinite loop to reuse processes with different data type | 13 November 2017, 10:54:50 UTC |
| 2d835b9 | ZHG | 13 November 2017, 10:53:50 UTC | Test program for the mpi parallelization function pfgemv_mpi added | 13 November 2017, 10:53:50 UTC |
| bdb57a8 | ZHG | 08 November 2017, 09:05:31 UTC | Prototype for hybride parallelization of fgemv added | 08 November 2017, 09:05:31 UTC |
| bfcde3f | ZHG | 07 November 2017, 13:33:34 UTC | MPI parallelization for fgemv but with a type inconsistency problem | 07 November 2017, 13:33:34 UTC |
| b25dd50 | ZHG | 07 November 2017, 13:25:34 UTC | Distributed memory parallelization of fgemv added Note: There exists type inconsistency for distribution of data with different types | 07 November 2017, 13:25:34 UTC |
| 2910102 | ZHG | 16 October 2017, 09:02:10 UTC | Merge branch 'parallelization-fgemv' of github.com:linbox-team/fflas-ffpack into parallelization-fgemv | 16 October 2017, 09:02:10 UTC |
| b39eccb | Clement Pernet | 16 October 2017, 08:58:00 UTC | fix bugs | 16 October 2017, 08:58:00 UTC |
| c02a1ce | Clement Pernet | 16 October 2017, 08:50:42 UTC | clean up | 16 October 2017, 08:50:42 UTC |
| b1f990d | ZHG | 13 October 2017, 14:59:37 UTC | Merge branch 'parallelization-fgemv' of github.com:linbox-team/fflas-ffpack into parallelization-fgemv | 13 October 2017, 14:59:37 UTC |
| 0ff2493 | ZHG | 13 October 2017, 14:54:55 UTC | First implementation using MPI on a ring topoloy of process added Further generalization and future improvement are on its way | 13 October 2017, 14:54:55 UTC |
| bdb4eea | Clement Pernet | 10 October 2017, 12:57:02 UTC | simplify pfgemv | 10 October 2017, 12:57:02 UTC |
| 5422e2b | Clement Pernet | 10 October 2017, 10:39:39 UTC | use modularbalanced double and display gfops | 10 October 2017, 10:39:39 UTC |
| 359752f | ZHG | 09 October 2017, 11:59:17 UTC | Performance for the parallel implementation of fgemv using row cutting strategy and threads as strategy parameter has been improved | 09 October 2017, 11:59:17 UTC |
| ea51123 | ZHG | 04 October 2017, 15:22:43 UTC | clean up make rule | 04 October 2017, 15:22:43 UTC |
| 25e9e4f | ZHG | 04 October 2017, 15:22:14 UTC | Clean up make rules | 04 October 2017, 15:22:14 UTC |
| 964e8b1 | ZHG | 04 October 2017, 14:02:02 UTC | Compile error for the test-pfgemv.C file caused by the incorrect usage of FFPACK::chooseField has been corrected | 04 October 2017, 14:02:02 UTC |
| f600e57 | ZHG | 04 October 2017, 13:29:54 UTC | Merge branch 'master' into parallelization-fgemv | 04 October 2017, 13:29:54 UTC |
| 93f6d07 | ZHG | 04 October 2017, 13:27:29 UTC | Merge branch 'parallelization-fgemv' of github.com:linbox-team/fflas-ffpack into parallelization-fgemv | 04 October 2017, 13:27:29 UTC |
| 0f30762 | ZHG | 04 October 2017, 13:26:27 UTC | Defaut sequential choice for the benchmark test of parallel implementation of fgemv added | 04 October 2017, 13:26:27 UTC |
| a5c01a7 | ZHG2017 | 04 October 2017, 12:57:58 UTC | Merge branch 'master' into parallelization-fgemv | 04 October 2017, 12:57:58 UTC |
| b5f4bd3 | ZHG | 04 October 2017, 12:41:37 UTC | Merge branch 'master' of github.com:linbox-team/fflas-ffpack | 04 October 2017, 12:41:37 UTC |
| adc34d2 | Jean-Guillaume Dumas | 04 October 2017, 09:51:36 UTC | no need to multiply zero matrices (and launches error in rns double in issue #64) | 04 October 2017, 09:51:36 UTC |
| 585641b | ZHG | 04 October 2017, 08:40:28 UTC | Benchmark for parallel implementation of fgemv named as pfgemv added | 04 October 2017, 08:40:28 UTC |
| 4a1f74d | ZHG | 03 October 2017, 11:32:54 UTC | Compile option for parallel benchmark of fgemv in Makefile.am added | 03 October 2017, 11:32:54 UTC |
| 44459b3 | ZHG | 03 October 2017, 11:30:47 UTC | Test program for parallel implementation of fgemv added | 03 October 2017, 11:30:47 UTC |
| 4a3a2d2 | ZHG | 03 October 2017, 11:26:01 UTC | Benchmark for parallel implementation of fgemv added | 03 October 2017, 11:26:01 UTC |
| deda204 | ZHG | 03 October 2017, 11:23:29 UTC | Parallel implementation of function fgemv using row cutting strategy and grain strategy parameter rewritten as required | 03 October 2017, 11:23:29 UTC |
| 4d074a3 | ZHG | 29 September 2017, 11:08:24 UTC | Test program for parallel implementation of fgemv added | 29 September 2017, 11:08:24 UTC |
| 5d526cc | ZHG | 29 September 2017, 11:05:15 UTC | Parallel implmentation of fgemv function using tile as grain size has been improved | 29 September 2017, 11:05:15 UTC |
| d9c0c38 | Clement Pernet | 28 September 2017, 14:24:15 UTC | remove typo | 28 September 2017, 14:24:15 UTC |
| 102e01d | ZHG | 28 September 2017, 13:44:19 UTC | Recursive parallel implementation of fgemv updated | 28 September 2017, 13:44:19 UTC |
| b9af1e9 | ZHG | 28 September 2017, 11:45:13 UTC | Cleaned up the pfgemv(parallel implementation of function fgemv) | 28 September 2017, 11:45:13 UTC |
| ea5d97d | ZHG | 28 September 2017, 11:43:50 UTC | Test program for parallel implementation of function fgemv added | 28 September 2017, 11:43:50 UTC |
| 4eaf1a7 | ZHG | 28 September 2017, 08:39:04 UTC | Cache-friendly parallel implementation of pfgemv using using blocking with tile as grain | 28 September 2017, 08:39:04 UTC |
| 5df36ec | ZHG | 27 September 2017, 09:13:01 UTC | Draft parallel version of fgemv using Grain as strategyParameter with cache friendly potential implementation added | 27 September 2017, 09:13:01 UTC |
| d2dab46 | ZHG | 26 September 2017, 13:13:59 UTC | pfgemv using row cutting strategy is corrected as required | 26 September 2017, 13:13:59 UTC |
| 960d136 | ZHG | 26 September 2017, 13:07:25 UTC | Correction of the pfgemv using recursive cutting strategy with thread as strategy parameter as required | 26 September 2017, 13:07:25 UTC |
| 60d4c42 | Clement Pernet | 26 September 2017, 12:15:13 UTC | move in jenkins maker | 26 September 2017, 12:15:13 UTC |
| 1c259b7 | Clement Pernet | 25 September 2017, 18:34:10 UTC | update output | 25 September 2017, 18:34:10 UTC |
| 2302dc4 | Clement Pernet | 25 September 2017, 17:56:45 UTC | fix output | 25 September 2017, 17:56:45 UTC |
| e4f645b | Clement Pernet | 25 September 2017, 15:22:16 UTC | fix bug in the test-suite with 32 bit archs | 25 September 2017, 15:22:16 UTC |
| ddd3c95 | ZHG | 25 September 2017, 14:26:50 UTC | Having Corrected some corner-case error for the parallel implementation of fgemv function: (1) with CuttingStrategy::Recursive and StrategyParameter::Threads (2) with CuttingStrategy::Recursive, StrategyParameter::Threads | 25 September 2017, 14:26:50 UTC |
| 86dcf06 | Clement Pernet | 25 September 2017, 14:10:22 UTC | fix output | 25 September 2017, 14:10:22 UTC |
| 77679a9 | Clement Pernet | 25 September 2017, 13:56:58 UTC | add missing include | 25 September 2017, 13:56:58 UTC |
| 98bc82f | Clement Pernet | 25 September 2017, 09:49:07 UTC | fix fedora pkg config dir | 25 September 2017, 09:49:07 UTC |
| c6dcb17 | Clement Pernet | 25 September 2017, 08:27:51 UTC | fix fedora builds | 25 September 2017, 08:27:51 UTC |
| 7456111 | ZHG | 25 September 2017, 07:35:00 UTC | Change the pfgemv function using line as grain to a function using rectangle blocking as grain | 25 September 2017, 07:35:00 UTC |
| 9c6bcc0 | Clement Pernet | 22 September 2017, 15:43:40 UTC | remove non standard output | 22 September 2017, 15:43:40 UTC |
| 9f2b148 | Clement Pernet | 22 September 2017, 15:22:08 UTC | adding some missing seeds | 22 September 2017, 15:22:08 UTC |
| 2147c0c | Clement Pernet | 22 September 2017, 15:13:01 UTC | revert missing include | 22 September 2017, 15:13:01 UTC |
| d75445b | Clement Pernet | 22 September 2017, 15:11:18 UTC | Merge branch 'master' of https://github.com/linbox-team/fflas-ffpack | 22 September 2017, 15:11:18 UTC |
| c4daa8c | Clement Pernet | 22 September 2017, 14:47:35 UTC | remove unused var | 22 September 2017, 14:47:35 UTC |
| a49fd6c | Clément Pernet | 21 September 2017, 09:29:38 UTC | Merge pull request #131 from linbox-team/seed_chooseField chooseField now uses seeds. update the test suite accordingly. | 21 September 2017, 09:29:38 UTC |
| 8f674ce | ZHG | 20 September 2017, 13:09:22 UTC | Take into considertation of the corner case of splitting matrix with StrategyParameter as Grain while the grain size is not a factor of the matrix dimension. | 20 September 2017, 13:09:22 UTC |
| b332a45 | ZHG | 20 September 2017, 12:10:09 UTC | Add a flexible way to split the matrix into fine grains for the parallel fgemv function | 20 September 2017, 12:10:09 UTC |
| 6577ad0 | ZHG | 19 September 2017, 15:01:04 UTC | Correct the wrong input parameter for the use of fdot inside pfgemv | 19 September 2017, 15:01:04 UTC |
| a85ee10 | ZHG | 19 September 2017, 14:49:49 UTC | Modification of pfgemv for the case of CuttingStrategy:=Recursive and StrategyParameter:=Thread Use fdot to compute the product of one line from the original matrix and the vector within one assigned thread | 19 September 2017, 14:49:49 UTC |
| 539c0da | ZHG | 19 September 2017, 12:44:51 UTC | Correct implementation error for the cuttingStrategy:row | 19 September 2017, 12:44:51 UTC |
| ca57014 | ZHG | 18 September 2017, 08:07:51 UTC | Program for unit test of pfgemv added into the directory of tests | 18 September 2017, 08:07:51 UTC |
| f616d97 | ZHG | 18 September 2017, 07:58:29 UTC | Parallelization of function fgemv as pfgemv in the sope of namespace FFLAS with following strategies: 1. CuttingStrategy: 'Recursive', StrategyParameter: 'Thread' 2. CuttingStrategy: 'Row', StrategyParameter: 'Thread' 3. CuttingStrategy: 'Row', StrategyParameter: 'Grain' | 18 September 2017, 07:58:29 UTC |
| e5c9865 | Clement Pernet | 15 September 2017, 14:35:32 UTC | clean binaries after run | 15 September 2017, 14:35:32 UTC |
| 3d50fb0 | Clement Pernet | 15 September 2017, 14:15:58 UTC | use atlas on fedora CI | 15 September 2017, 14:15:58 UTC |
| 68f09f6 | Clement Pernet | 15 September 2017, 13:46:46 UTC | Merge branch 'master' of https://github.com/linbox-team/fflas-ffpack | 15 September 2017, 13:46:46 UTC |
| 0624c77 | Clement Pernet | 15 September 2017, 13:46:32 UTC | update jenkins maker to new slaves | 15 September 2017, 13:46:32 UTC |
| 48430d8 | Clément Pernet | 15 September 2017, 10:52:37 UTC | Merge pull request #133 from linbox-team/ftrtri_implem fix interface for ftrtri | 15 September 2017, 10:52:37 UTC |
| a16a835 | Clement Pernet | 15 September 2017, 10:02:53 UTC | fix interface for ftrtri | 15 September 2017, 10:02:53 UTC |
| b324d89 | ZHG | 15 September 2017, 09:39:20 UTC | Add test source code for pfgemv | 15 September 2017, 09:39:20 UTC |
| 73b2ea0 | Clement Pernet | 14 September 2017, 15:23:38 UTC | update detection of clang, fix bug with clang38 not detecting __float128 and add explicit link to libpthreads | 14 September 2017, 15:23:38 UTC |
| 111b15b | Clément Pernet | 13 September 2017, 15:44:35 UTC | Merge pull request #125 from linbox-team/test-solve Test solve | 13 September 2017, 15:44:35 UTC |
| f5bc3d8 | Clement Pernet | 13 September 2017, 11:21:08 UTC | prettier output | 13 September 2017, 11:21:08 UTC |
| d60611f | Clement Pernet | 13 September 2017, 10:01:52 UTC | fix typo | 13 September 2017, 10:01:52 UTC |
| 02f4438 | ZHG | 13 September 2017, 09:06:10 UTC | Replace Field::Element with Field::Element_ptr | 13 September 2017, 09:06:10 UTC |
| d5c61b4 | ZHG | 13 September 2017, 08:41:52 UTC | Clean up the conflict after Merging branch 'test-solve' of github.com:linbox-team/fflas-ffpack into test-solve | 13 September 2017, 08:41:52 UTC |
| 31e0b88 | ZHG | 13 September 2017, 08:15:34 UTC | Clean up the code with given comments | 13 September 2017, 08:15:34 UTC |
| 16c7497 | Clement Pernet | 12 September 2017, 13:08:10 UTC | fix merge error messages | 12 September 2017, 13:08:10 UTC |
| 669fc62 | Clement Pernet | 12 September 2017, 11:58:23 UTC | fix generation of the system's matrix | 12 September 2017, 11:58:23 UTC |
| 19ecb83 | Clement Pernet | 08 September 2017, 12:41:33 UTC | fix bug in Solve | 08 September 2017, 12:41:33 UTC |
| 1b2a84e | Clement Pernet | 07 September 2017, 17:41:49 UTC | chooseField now uses seeds. update the test suite accordingly. | 07 September 2017, 17:41:49 UTC |
| 4277600 | Clement Pernet | 07 September 2017, 11:22:43 UTC | second thought: better to clear the charpoly list when failure is caught. | 07 September 2017, 11:22:43 UTC |
| 100fcaa | Clément Pernet | 06 September 2017, 15:40:55 UTC | Merge pull request #130 from linbox-team/bug_charpoly Bug charpoly | 06 September 2017, 15:40:55 UTC |
| 53572e1 | Clement Pernet | 06 September 2017, 14:28:50 UTC | revert other tests temporarily removed for debugging | 06 September 2017, 14:28:50 UTC |
| 5882e13 | Clement Pernet | 06 September 2017, 14:27:33 UTC | clean up | 06 September 2017, 14:27:33 UTC |
| 0ccff85 | Clement Pernet | 06 September 2017, 14:26:09 UTC | remove unecessary include | 06 September 2017, 14:26:09 UTC |
| 91ab5ca | Clement Pernet | 06 September 2017, 14:22:21 UTC | fix rare bug: when probabilistic algorithm fails, the list of invariant factors was not emptied and polluted the ones computed in the next attempt. | 06 September 2017, 14:22:21 UTC |
| 59e40c2 | Clement Pernet | 06 September 2017, 08:46:03 UTC | Merge branch 'master' into test-solve | 06 September 2017, 08:46:03 UTC |
| 6a7ab49 | Clement Pernet | 06 September 2017, 08:41:41 UTC | tracking the bug | 06 September 2017, 08:41:41 UTC |
| de55c19 | Clément Pernet | 05 September 2017, 12:09:59 UTC | Merge pull request #126 from linbox-team/solve-tutorial Tutorial for the solve function and the fgesv function | 05 September 2017, 12:09:59 UTC |
| 5bddf43 | Clement Pernet | 04 September 2017, 11:10:22 UTC | typo | 04 September 2017, 11:10:22 UTC |
| e90f430 | Clément Pernet | 04 September 2017, 09:53:19 UTC | Merge pull request #127 from linbox-team/pfdot fix bug when GIVARO does not use OMP | 04 September 2017, 09:53:19 UTC |
| 3a65b89 | Clement Pernet | 04 September 2017, 09:50:46 UTC | fix bug when GIVARO does not use OMP | 04 September 2017, 09:50:46 UTC |
| 06ed360 | ZHG | 04 September 2017, 09:24:44 UTC | Tutorial for the solve function and the fgesv function | 04 September 2017, 09:24:44 UTC |
| 44e5310 | ZHG | 01 September 2017, 11:20:18 UTC | Correction for the use of coefficient array for the function Solve | 01 September 2017, 11:20:18 UTC |
| 732fe3c | ZHG | 01 September 2017, 11:18:54 UTC | Correction for the tutorial program of using function solve | 01 September 2017, 11:18:54 UTC |
| ecb881c | ZHG | 01 September 2017, 11:15:06 UTC | The last checked test prgram for the function solve | 01 September 2017, 11:15:06 UTC |
