chanwit wrote:
John,
The papers said about generating bytecodes from what they called
'trace'. The representation is their trace is a kind of SSA variant.
When the system semantic is altered, they re-record or extend the trace.
Each time they got a complete trace for each method, the trace will be
compiled into the machine language (in our case JVML).
It looks good enough for me to try. I'll write some small code to
prove its concept. Whether it works or not, I will post here ;-)
The thing is we already have a very sophisticated JIT as part of the JVM. It
will do a really excellent job if you let it. If you keep messing with the
bytecodes you will not allow the JVM JIT to do its optimisation.
The trick of getting a fast dynamic runtime seems to be to do as little
optimisation as possible just have short simple execution paths and then let
the JVM JIT do the work.
John Wilson
--
View this message in context:
http://www.nabble.com/as-usual-on-performance---operations-on-primitives-tp15713441p15717677.html
Sent from the groovy - dev mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
31 messages in org.codehaus.groovy.devRe: [groovy-dev] as usual on performa...
