DD-AVX

DD-AVX: Library of high-precision operations accelerated by AVX.

author's page
http://www.slis.tsukuba.ac.jp/~s1530534/index.html
e-mail
hishinuma@slis.tsukuba.ac.jp

Double-Double (DD) precision operations are used to reduce rounding errors and improve the convergence of Krylov subspace methods.

This library has Double-Double precision operations accelerated by AVX and AVX2.
AVX and AVX2 are intel SIMD instructions. They operate four double precision operation simultaneously.

This library needs to merge "Lis" for DD precision iterative solvers.
Official support "Lis" version is "lis-1.4.58."
After merge "lis" and "DD-AVX," the interface is same as "lis."

Lis web page
http://www.ssisc.org/lis/index.en.html

This library is managed by git. old version can search "code" page.