|Compiler Synthesized Integer Division. napi@rangkom.MY (1991-12-14)|
|From:||napi@rangkom.MY (Mohd Hanafiah Abdullah)|
|Date:||14 Dec 91 09:31:54 GMT|
With regard to the 1987 ACM paper entitled "Integer Multiplication and
Division on the HP Precision Architecture" by Daniel J. Magenheimer, et al,
[in the 1987 ASPLOS proceedings, pp. 90-99 -John]
could someone please explain the DIVISION part in "simpler" and "clearer"
way(s)? Given that you have read it, of course. For example, how does one
get the Z value?
On the same subject, I wonder if there is a Z value that is applicable to all
situations for 32 bit divisions. From the shown examples, 2 to the power of
36 is the highest for small odd divisors. Are there higher ones for some
other cases? Excuse me for being naive!
Please respond by email.
Return to the
Search the comp.compilers archives again.