freenode/#clasp - IRC Chatlog
Search
12:01:48
kpoeck
Finally was able to reproduce the problem that (core:source-pos-info-function-scope spi) returns nil.
12:02:49
kpoeck
When the compiled-file of the following is loaded https://github.com/sharplispers/ironclad/blob/master/src/common.lisp#L52
18:27:26
drmeister
I'm trying to do a simple common chemical operation - compare the position of every atom to every other atom for a moderately large system of atoms - N^2 comparisons - that's 1Billion pairwise comparisons will take 20 minutes with Cando as it is currently setup.
18:29:17
drmeister
That's doing it the naive way, putting all of the atoms into a vector, looping over the atoms with an inner loop over the rest, pulling up the position of each atom and calculating the distance.
18:51:14
drmeister
The reason I bring this up is there are times when the naive approach would be very quick to implement but it's slow now because doing anything with double precision values or getting a vector conses memory. That slows things way, way down.
20:27:05
Lord_Nightmare
drmeister: sounds similar to what elmar krieger's YASARA thing does, except i don't know if that is lisp/clasp based
20:42:32
drmeister
Except they don't appear to be based an accepted molecular dynamics engine (AMBER in our case).
20:45:00
drmeister
Written in something called "Portable Vector Language" - do you know what that is?
20:55:28
Bike
i see a Scientific Vector Language used for computational chemistry, but it's proprietary
21:02:03
drmeister
They say they wrote YASARA in PVL - and there is little indication what PVL is. Weird.