GSS_spd 2.2 发布

GSS_spd is a novel sparse solver for symmetric positive definite matrices (SPD in abbreviation).
GSS_spd is about 20% faster than PARDISO in numerical factorization. The result is based on nearly
all the large SPD matrices (with more than 300,000 unknowns) from UF sparse matrix collection,
which are indisputably authoritative and representative.
The testing CPU is Intel Q8200, which is cheap now. And the average numerical factorization time is less than 10 seconds!
So it’s easy and fast to solve million SPD matrices in PC.

 

A free version for 64 bit windows is available at http://www.grusoft.com/GSS_spd.htm.