You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Now that Elemental supports searching for (integer) relations using LLL and/or BKZ, it would make sense to implement the HJLS and/or PSLQ algorithms. There is some contention as to whether or not these algorithms are distinct [1], but it would appear that the latter is more accurate when using finite-precision arithmetic.
Also, for the sake of making the historical connection, the need for higher-precision within PSLQ appears to be one of the reasons that Bailey et al. wrote their higher-precision arithmetic packages (with QD also incorporating techniques from the code of Briggs and Kahan).
benson31
added a commit
to benson31/Elemental
that referenced
this issue
Dec 15, 2021
* Add cuSOLVER-based implementation of Hermitian eigensolves.
* Add rocSOLVER implementation of Syev.
* Reenable the HermitianEig test for sequential matrices
a0x8o
added a commit
to a0x8o/elemental
that referenced
this issue
Jun 6, 2022
* LLNL-hydrogen: (711 commits)
Update README.md
Add Clang 4.0 to .travis.yml and allow 6.0 to fail
Fix ambiguous template special. errors with GCC 7+
Add GCC 8 and Clang 6.0 to Travis CI.
Enable OpenMP code only if EL_HYBRID is set
CMake, GCC: Enable libquadmath detection for GCC 5.x to 8.x.
Pass MPI_C_COMPILER instead of CMAKE_C_COMPILER to ensure that ParMetis is built with the right compiler (flags)
Cleanup some issues with ETI and the Memory class (elemental#136)
Distributed Cholesky for GPU (elemental#135)
Debugging TranslateBetweenGrids (elemental#133)
Fix an issue leading to bad linkage. (elemental#134)
Update CMake support for CUDA HPC SDK and HIP (elemental#132)
Update HydrogenConfig.cmake.in
Update CMakeLists.txt
Update CMakeLists.txt
Add assertion that LAPACK is found. (elemental#131)
Serial eigensolver for GPU (elemental#130)
Remove FP16 from the ETI for eigensolvers. (elemental#129)
Changed the CUDA compiler detection to just enable CUDA support rather (elemental#128)
Reenable the HermitianEig API for sequential matrices. (elemental#127)
...
Now that Elemental supports searching for (integer) relations using LLL and/or BKZ, it would make sense to implement the HJLS and/or PSLQ algorithms. There is some contention as to whether or not these algorithms are distinct [1], but it would appear that the latter is more accurate when using finite-precision arithmetic.
[1] http://perso.ens-lyon.fr/damien.stehle/downloads/PSLQHJLS.pdf
The text was updated successfully, but these errors were encountered: