Re: LW compiler optimizations
"Yuri Davidovsky (as work at disclosure dot ie)" <[email protected]>
| Newsgroups | gmane.lisp.lispworks.general |
|---|---|
| Message-ID | <[email protected]> |
> On 15 Jan 2025, at 19:40, David McClain <[email protected]> wrote: > > A seeming small change in Lisp structure can make a big change to overall timings I remember having a similar slowdown when I tried your trick of using a single element typed array with just one double float in it as an accumulator, however it actually made it slower versus using just a double float type r declared in let. It should have worked at least the same fast, but I digress. Now, from my estimates of 3 cycles per a pair operation in native lisp code it does not appear to be possible to use a boxed double float as an accumulator, so it appears as if the compiler found a way to keep the double float in a register as a raw datum. _______________________________________________ Lisp Hug - the mailing list for LispWorks users [email protected] http://www.lispworks.com/support/lisp-hug.html