Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
MEGGITT's algorithms
8 views
Skip to first unread message
Baykov Vladimir
May 16, 1995, 3:00:00 AM5/16/95
to
I suggest to pay attention of the people in Computer Arithmetic field
to the paper of J.E.Meggitt, published in 1962 in IBM J. Res &Dev.
Dr. Meggitt was the first who suggested the algorithms similar to CORDIC,
but using normalised constants and normalised variables reducible to zero.
It allowed him to improve significantly the accuracy of the iterative al-
gorithms.
In addition relative errors of the fixed-point representation numbers in
Meggitt's algorithms in contrast with Volder's algorithms practicaly
don't depend on the word size, like in the floating-point number
representation.
In other words, Dr.Meggitt suggested not only "pseudodivision and pseudo-
multiplication processes" (as indicated in the title of his paper), but also
"pseudofloating- point number representation" for CORDIC method.
All who are interested in details I can suggest to find in "Computer and
Control Abstracts", Author Indexes, beginning from 1971, all my papers
related to this topic.My last name: BAYKOV (sometimes written: Baikov).
to the paper of J.E.Meggitt, published in 1962 in IBM J. Res &Dev.
Dr. Meggitt was the first who suggested the algorithms similar to CORDIC,
but using normalised constants and normalised variables reducible to zero.
It allowed him to improve significantly the accuracy of the iterative al-
gorithms.
In addition relative errors of the fixed-point representation numbers in
Meggitt's algorithms in contrast with Volder's algorithms practicaly
don't depend on the word size, like in the floating-point number
representation.
In other words, Dr.Meggitt suggested not only "pseudodivision and pseudo-
multiplication processes" (as indicated in the title of his paper), but also
"pseudofloating- point number representation" for CORDIC method.
All who are interested in details I can suggest to find in "Computer and
Control Abstracts", Author Indexes, beginning from 1971, all my papers
related to this topic.My last name: BAYKOV (sometimes written: Baikov).
0 new messages