[XSL-LIST Mailing List Archive Home]
[By Thread]
[By Date]
Re: Java performance (Saxon)
Subject: Re: Java performance (Saxon) From: Niclas Hedhman <niclas@xxxxxxxxxxxx> Date: Wed, 01 Mar 2000 10:38:08 +0800 |
Kay Michael wrote: > A B C D E > Saxon 5.2, Sun 9.4 15.9 19.3 20.1 5.3 > Saxon 5.2, MS 4.9 9.5 11.8 12.9 20.0 > Saxon 5.3, Sun 7.3 12.3 18.4 18.0 5.8 > Saxon 5.3, MS 3.1 5.2 19.0 10.8 19.8 Is this with the JIT enabled?? I bet it was. Very common to see these kind of results. What happens if you turn it off? I believe that the figures will be much more stable, and pointing in the same direction. JIT compilation has a tendency to have various speed improvements. And there is also the fact that the compilation takes time, and I believe that the whole chunk is not compiled in one shot, meaning that when you have introduced new code, one JIT compiler might have some additional overhead for that section while the other has not. And Bill, HotSpot would probably barely get started by the time the execution is over ;o) Niclas XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: Java performance (Saxon), Jon Smirl | Thread | asp problem, Edmond Bramhall |
Re: Java performance (Saxon), Jon Smirl | Date | RE: Repetition without a repeated s, Mike Brown |
Month |