Search
1:04:36
ERIIXBlaike[m]
** NICK eriix
4:53:21
drmeister
The problem is double precision math is still all boxed.
4:53:32
drmeister
I rearranged the code and made it about 10x faster.
4:53:58
drmeister
Then I implemented the inner loop in C++ and exposed it - that was at least another 10x faster again.
4:54:12
drmeister
On to the next bottleneck
5:13:09
beach
Good morning everyone!
13:00:56
kpoeck
Regarding compile-file-parrallel
13:01:09
kpoeck
Looking at https://github.com/clasp-developers/clasp/blob/dev/src/lisp/kernel/cmp/compile-file-parallel.lsp#L153