Reviews and Descriptions of Tables and Books

1956 Mathematics of Computation  
studied by Todd, for which 7?(vl") ~ c-n2-2~in for a certain constant c. No direct comparison of the results of our experiments with those made by Todd [1] (for a 49 X 49 matrix, without iteration) is possible, since his results are expressed in terms of the norm of the error matrix. The calculating program thus described operates at a fixed binary point and uses 40 binary digits. A program operating at a floating binary point and including 32 binary digits in its digital portion and 8 binary
more » ... gits in its scalar exponent has also been developed. This program works at an approximately 30% slower rate and gave, when applied to the above-mentioned systems, a slightly better solution. Hans Riesel Essingeringen 9 Stockholm Sweden 1. These are radix tables with three figure groups; when using the decimal to octal table the tabular entries are added together using octal arithmetic. The decimal to octal table has 11 place octal values of n X 10~K for n = 0(1)999 and K = 3, 6, 9. The octal to decimal table has 10 place decimal values of « X 8~x for n = 0(1)511 and K = 3, 6, 9, 12, together with an argument column giving n X 8-3 in octal. The tables were printed by photo-offset from output of an IBM tabulator. This machine has the unpleasant characteristic of leaving more space between adjacent columns than rows. The tables were developed for use with the RAYDAC.
doi:10.1090/s0025-5718-56-99306-1 fatcat:hg3tjkqavrbczc3jeyfsjaaalq