Re: More on Lisp structures performance
"David McClain (as dbm at refined-audiometrics dot com)" <[email protected]>
| Newsgroups | gmane.lisp.lispworks.general |
|---|---|
| Message-ID | <[email protected]> |
Yes, IIRC, it was a note from Martin that described the packed format for SP float. > On Jan 22, 2026, at 12:29, Yuri Davidovsky (as work at disclosure dot ie) <[email protected]> wrote: > > >> On 22 Jan 2026, at 18:06, David McClain <[email protected]> wrote: >> >> Ah yes, I recall the previous discussion. Your first response paragraph fills in the puzzle. > > Actually we haven’t discussed that, the “another email” is the one I am planning to write soon that will go into the technical details how I optimised the serialised structs access. It was a bit of a rabbit hole but it had an end. I did not get it to run as fast as it theoretically should but in return I got a good understanding how the LW compiler behaves when generating computational code and why SBCL is able to generate native performance output. > > I am not even sure actually that it uses vectorisation all that much, there is a lot lower hanging fruit. > > _______________________________________________ > Lisp Hug - the mailing list for LispWorks users > [email protected] > http://www.lispworks.com/support/lisp-hug.html _______________________________________________ Lisp Hug - the mailing list for LispWorks users [email protected] http://www.lispworks.com/support/lisp-hug.html