An efficient sparse matrix multiplication scheme for the CYBER 205 computer
Scientific and Technical Information Division, National Aeronautics and Space Administration, 1988 - Computers - 8 pages
What people are saying - Write a review
We haven't found any reviews in the usual places.
1.0 Density Figure 64-bit words needed algorithm bit pattern bit vector central processing unit Compressed vector Computational Requirements computational scheme Control Data CPU and storage CPU require CYBER 205 computer dense diagonal are stored diagonal i-1 st diagonal ith diagonal diagonal length diagonal storage diagonal type diagonal with length diagonal-based Efficient Sparse Matrix Estimates the CPU expand finite element formulations FORTRAN fraction of nonzeros goals of minimizing i-1 st diagonal index vector initialization subroutine Langley Research Center main diagonal Matrix Multiplication Scheme MFLOPS NASA nonzeros are stored nonzeros lie nsec per element number of 64-bit number of nonzeros pand percent Performing Organization putation Q4CMPYD serial computation sparse and long sparse and short Sparse Matrix Multiplication st diagonal ith storage and computational storage estimates storage requirement storage type subdiagonal subroutine Q4CMPCTD superdiagonal sw and cw type 2 diagonal Type 2 Type user-provided weights vector computation vector plus row weighted resource requirement zeros