Thanks, Marty. (And Thorbjørn and Joe Sam for your responses too.) I tried an experiment in Qshell using a jar file with no JVAPGM attached:
java -Djava.compiler=jitc -cp ...
and the first line in stdout was
Attaching Java program to /java/edi/jars/edi.jar.
So there's still a minimal JVAPGM being generated, with OPTIMIZATION(*INTERPRET). Same thing happened when I used "jitc_de" as the value. But now that I read that page you linked to more carefully, it does say "All modes verify the code and create a Java program object to hold the preverified form of the program." (As Thorbjørn said.) So it looks like I can't avoid the built-in CRTJVAPGM step no matter what I do.
And the answer to my original question, why does it take so long, appears to be that it takes a ridiculously long time on machines where the jar file is in a journaled directory. We're talking to our HA supplier right now to find out what's up with that.
From: java400-l-bounces@xxxxxxxxxxxx [mailto:java400-l-bounces@xxxxxxxxxxxx] On Behalf Of Urbanek, Marty
Sent: September 26, 2007 08:39
Subject: RE: Why does CRTJVAPGM take so long sometimes?
Have you tried setting Java system property java.compiler to "jitc" to make it use the JIT compiler?
I'm not saying what the results should be, but IBM has been saying for the last couple of releases that JIT should be as fast or faster than direct executable (what comes out of CRTJVAPGM).
Here's a link:
As an Amazon Associate we earn from qualifying purchases.