freenode/#sbcl - IRC Chatlog
Search
11:18:16
flip214
Using (TIME ...) gives me 1.1sec (real or user). Using sb-sprof:with-profiling tells me "sampling time: 0.0279 sec" for the same form.
11:19:12
flip214
why doesn't sprof trace for longer? I've got :sample-interval 0.0001 and :max-samples 1000000
14:23:24
_death
this paper has a useful glossary https://www.researchgate.net/publication/221252239_Python_compiler_for_CMU_common_Lisp
14:48:13
_death
hmm, seems there's a more up-to-date paper, https://common-lisp.net/project/cmucl/downloads/doc/CMUCL-design.pdf (via https://github.com/guicho271828/sbcl-wiki/wiki )
15:35:39
lukego
please don't say "more up to date" when talking about a 100 page document that I have recently printed hardcopy ;-)
15:36:46
lukego
thanks. yeah I mostly just wondered about the acronym so I don't guess constantly while reading
15:38:29
stassats
although, it doesn't really have a name, and not sure what's so temporary about it
15:40:28
lukego
what would really help right now is some handy REPL functions for printing the various IR representations of simple examples like (DEFUN FOO (X) (1+ x))
15:40:50
lukego
are those easy to call interactively? or how do you see what the compiler is thinking?
15:43:51
stassats
can never get around to creating the former, as the latter is usually enough, even if i have to insert it somewhere in the middle
15:47:59
stassats
because the answer to it is immaterial - you shouldn't ever have a defpackage without :use