libsharp2/README.md
2021-05-20 12:38:51 +00:00

45 lines
1.6 KiB
Markdown

NOTICE
======
Active development of this package has stopped. The package will receive bug fixes
if necessary, but otherwise the code has been integrated into the `ducc0`
package (https://gitlab.mpcdf.mpg.de/mtr/ducc), and further development is
taking place there.
Please prefer `ducc0` over `libsharp2` if you are starting a new project!
# Libsharp2
Library for efficient spherical harmonic transforms at arbitrary spins,
supporting CPU vectorization, OpenMP and MPI.
## Paper
https://arxiv.org/abs/1303.4945
## News
### January 2019
This update features significant speedups thanks to important algorithmic
discoveries by Keiichi Ishioka
(https://www.jstage.jst.go.jp/article/jmsj/96/2/96_2018-019/_article and
personal communication).
These improvements reduce the fraction of CPU time spent on evaluating the
recurrences for Y_lm coefficients, which means that computing multiple
simultaneous SHTs no longer has a big performance advantage compared to SHTs
done one after the other.
As a consequence, libsharp's support for simultaneous SHTs was dropped, making
its interface much simpler.
With the proper compilers and flags (see the file COMPILE for details) libsharp2
is now built with support for SSE2, AVX, AVX2, FMA3, FMA4 and AVX512f and the
appropriate implementation is selected dynamically at runtime. This should
provide a very significant performance boost for everyone using pre-compiled
portable binaries.
## Compilation
The library uses the standard `autotools` mechanism for configuration,
compilation and installation. See the file `COMPILE` for configuration hints.