AMUXArithmetic Logic Unit Multiplexer
AMUXAvionics Multiple Bus
Copyright 1988-2018, All rights reserved.
References in periodicals archive ?
In Table II, we compare the execution time of the automatically generated general sparse columnwise version of SGEMV for [Alpha] = 1 and [Beta] = 0 with the execution time of the hand-coded (but not-annotated) routine AMUX of SPARSKIT [Saad 1990] (computing y [left arrow] Ax) for a few matrices from the Harwell-Boeing Sparse Matrix Collection [Duff et al.
Matrix n nnz CW-SGEMV steam2 600 13760 4.92 [multiplied by] [10.sup.-4] jpwh_991 991 6027 7.52 [multiplied by] [10.sup.-4] gre_1107 1107 5664 8.33 [multiplied by] [10.sup.-4] orani678 2529 90158 2.53 [multiplied by] [10.sup.-3] Matrix AMUX steam2 7.27 [multiplied by] [10.sup.-4] jpwh_991 1.1 [multiplied by] [10.sup.-3] gre_1107 1.23 [multiplied by] [10.sup.-3] orani678 3.16 [multiplied by] [10.sup.-3] 3.1.2 SGEMV: Banded Scheme and Alternative Operations.