JRuby on OpenVMS works! (slow, though)

[synergism]

I was pleasantly surprised that JRuby works on OpenVMS. Here's a
discussion I kicked off on the HP ITRC forums about performance
concerns:

http://forums12.itrc.hp.com/service/forums/questionanswer.do?threadId=1218567

Ben

 

[Dan]

[Note: parts of this message were removed to make it a legal post.]

Hi Ben

I've been messing with JRuby on VMS for a little while already.

I was suprised to see that jirb even runs inspite of the massive stacktrace
that occurs because it thinks it's running on a UnixTerminal.

Here's hoping that jruby will be a viable option for OpenVMS in the near
future.

 

[M. Edward (Ed) Borasky]

I was pleasantly surprised that JRuby works on OpenVMS. Here's a
discussion I kicked off on the HP ITRC forums about performance
concerns:

http://forums12.itrc.hp.com/service/forums/questionanswer.do?threadId=1218567

Ben

Yes, but does *OpenVMS* work?

 

[Charles Oliver Nutter]

Hi Ben

I've been messing with JRuby on VMS for a little while already.

I was suprised to see that jirb even runs inspite of the massive stacktrace
that occurs because it thinks it's running on a UnixTerminal.

Here's hoping that jruby will be a viable option for OpenVMS in the near
future.

Please file a bug for that...we definitely want all platforms where we
run to run well, so this sounds like a good one to fix

- Charlie

 

[Charles Oliver Nutter]

I was pleasantly surprised that JRuby works on OpenVMS. Here's a
discussion I kicked off on the HP ITRC forums about performance
concerns:

http://forums12.itrc.hp.com/service/forums/questionanswer.do?threadId=1218567

I scanned through the discussion, and do have a few suggestions:

- Try various JVM flags to adjust performance...I have no idea what
they'd be under VMS.
- Try JRuby flags, available by running jruby --properties. Several in
there, especially "reflection" might improve performance or startup time.
- If you have access to the OpenVMS JVM folks, they might be able to
point out things we're doing in JRuby that interfere with their Java
optimizations. For example, the IBM J9 guys warned us against re-using
the same block of code for exception handled and non-exception-handled
code paths, which breaks certain optimizations they do.

Performance is very important to us, so please file bugs for anything
concrete you're able to find.

http://jira.codehaus.org/browse/JRUBY

- Charlie