Lousy performance on FOR loop
|
|
Bugzilla Link |
75 |
Created on |
Oct 22, 2004 23:13 |
Resolution |
FIXED |
Resolved on |
Aug 09, 2006 10:14 |
Version |
1.00beta |
OS |
Windows NT |
Architecture |
PC |
Attachments |
bug75prd.sac, bug75.sac
|
Extended Description
I've compiled "plus reduce iota n" written as a FOR loop and
directly in sac, with n <- 5e6.
APL interpreter doing +/iota n: About 3 seconds CPU on Celeron creaky box.
sac doing +/iota n compiled from APL: About 1.27 seconds CPU
sac doing FOR loop version of same: 6.1-6.7 seconds real, about 5.9 seconds CPU.