The design of a parallel dense linear algebra software library: Reduction to Hessenberg, tridiagonal, and bidiagonal form (Q1904151): Difference between revisions

The paper discusses issues in the design of ScaLAPACK (`Scalable LAPACK'), a software library for performing dense linear algebra computations on distributed memory concurrent computers. The ScaLAPACK routines can be used for reducing matrices to Hessenberg tridiagonal, and bidiagonal forms. These routines are important in the solution of eigenproblems. Upon completion, ScaLAPACK will make available on distributed memory machines the same set of library routines that LAPACK provides for vector and shared memory architectures. The use of BLAS routines, their parallel versions (PBLAS), and the communication subprograms (BLACS) is discussed. There is a tradeoff between efficiency and software engineering considerations, such as ease of programming and simplicity of code. The implementation details, performance, and scalability of the ScaLAPACK factorization routines are presented in a separate paper. The ScaLAPACK reduction routines are currently available through netlib for all numerical data types, such as single and double precision real and complex. To obtain the routines, send the message `send index from scalapack' to netlib\@ornl.gov.