Hi folks,
lucky me I just got upgraded to a brand new supposedly powerfull machine a few
days ago. We are in the process of testing performance improvments. Following
you will find a script which ran perfectly fine on the older machine (350 Mz.
Pentium II with 256M of RAM and 2G of swap file) . It ran in 46h.11min. Under
the new machine (AP500 with 512 of RAM, 3G of swap file and dual processors) the
first part ran, i.e. the "stepAIC" function from the MASS library (7h10m vs
34h51m on the old machine) but the last function, i.e. "dropterm" again from the
MASS library, bombed with an error unknown to me.
Any clue(s) would be appreciated, thanks,
Here is the script
**************************************
ttt.time2 _ proc.time()
inc2.freq.glm.step2 _ stepAIC(object = inc2.freq.glm.min,
scope = inc2.freq.scope, scale = inc2.phi.freq, direction = 'both',
k = log(log(nrow(inc2.freq))))
inc2.freq.stepAIC.time2 _ ElapsedTime(ttt.time2, proc.time())
synchronize()
ttt.time2_proc.time()
inc2.freq.glm.step2.anova_inc2.freq.glm.step2$anova
inc2.freq.glm.step2.anovaF_anova(inc2.freq.glm.step2, test = 'F')
inc2.freq.glm.step2.dropterm_dropterm(inc2.freq.glm.step2,
scale = inc2.phi.freq, k = log(log(nrow(inc2.freq))), test = 'Chisq')
Error: Big file (C:\Program Files\sp2000\splus\_Functio/__BIG) has bad header
****************************************
Gérald Jean
Analyste-conseil (statistiques), Actuariat
télephone : (418) 835-8839
télecopieur : (418) 835-5865
courrier électronique: gerald.jean@spgdag.ca
"In God we trust all others must bring data"
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|