"AL" == Anders Lindback <[EMAIL PROTECTED]> writes: AL> Barnet Wagman skrev: >> >> By default, the Java compilers includes line numbers in the bytecode >> they produce. Is there much benefit to compiling production code >> without line numbers (-o option, I believe), in terms of speed?
AL> Removing the line number and other debugging info makes the AL> compiled byte code smaller and with a smaller memory footprint AL> an application may run faster. But only if you have a performance AL> problem where applicaitons are using too much memory. AL> (Performance problems usually are one of CPU speed, memory or I/O.) AL> In theory using the -O optimizer option should make the AL> code faster - but it all depends on the application AL> if you see any big change. I thought that line numbers were kept in the attribute section of the classfile, and that the bytecodes were identical. Is this not correct? Also, there is no documented -O option with Sun's 1.4.0 Java compiler (Search for "-O" on the javac documentation page, http://java.sun.com/j2se/1.4/docs/tooldocs/solaris/javac.html) although it is accepted and the command "javac -help" lists it. Compiling a 409 line Java source file produced no difference in bytecode between the "optimized" compile and the non-optimized compile. I believe that the rationale for removing the -O functionality is that mixed-mode interpreters like HotSpot do a much better job of (on-the-fly) optimization than static, compile-time optimizers can do. --- Vladimir -------- Vladimir G. Ivanovic http://leonora.org/~vladimir 2770 Cowper St. [EMAIL PROTECTED] Palo Alto, CA 94306-2447 +1 650 678 8014 AL> ---------------------------------------------------------------------- AL> To UNSUBSCRIBE, email to [EMAIL PROTECTED] AL> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] ---------------------------------------------------------------------- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
