Re: static estimation of conditional branches?

>[As I recall, Fortran II dropped FREQUENCY because it was infrequently
>used and made little difference. I've heard that it may even have been
>implemented backwards and nobody noticed. -John]


I was told the story about 10 years ago by, I think, Lois Nash, at
Yorktown Heights. She (or whoever told me the story if I've remembered
wrong) did the backend for the compiler that was to use these directives.
Sometime after releasing it, she took a vacation and someone reversed the
sense of what was passed to her in an earlier phase. About a year after
this happened, she was debugging some object code and noticed the
reversal. This was when people grovelled over the object code--in the
1950s, I guess--because of their general suspicion of compilers (*we*
never do that), yet no one ever pointed it out to her. She figures nobody
noticed.


(By the way, the Multiflow compiler allows programmers to insert
directives to this effect to guide the compiler in chosing long regions
within which to do code motions.)


-- Josh Fisher // HP Labs
--

Return to the comp.compilers page.
Search the comp.compilers archives again.