Some of you may recall that I urged caution in a response to Edward
Jaffe's post about the use of unsigned arithmetic.  (His specific
interest was in establishing mnemonics for the jumps employed with
unsigned binary-integer values.)

This morning, in the course of writing a BXLEG instruction in a code
skeleton, I had a minor epiphany.  BXH and BXLE use signed four-byte
binary-integer arithmetic, and BXHG and BXL:EG of course use signed
eight-byte binary-integer arithmetic.

Unsigned quantities are thus not immediately usable with indexed
branching instructions.    For the uses EJ contemplates, this is, I
imagine, unimportant.  For some of the computational work I do it is a
severe limitation.   I should have, but did not, mention this in my
original post.


John Gilmore, Ashland, MA 01721 - USA

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to