< prev index next >
test/java/text/Format/DecimalFormat/FormatMicroBenchmark.java
Print this page
*** 49,68 ****
* Note:
* - Vm options -Xms, -Xmx, -XX:NewSize must be set correctly for
* getting reliable numbers. Otherwise GC activity may corrupt results.
* As of jdk80b48 using "-Xms500m -Xmx500m -XX:NewSize=400m" covers
* all cases.
! * - Optionally using "-XX:+printGC" option provides information that
* helps checking any GC activity while benches are run.
*
* Vm Options:
* - Vm options to use (as of jdk80b48):
* fast-path case : -Xms128m -Xmx128m -XX:NewSize=100m
* non fast-path case: -Xms500m -Xmx500m -XX:NewSize=400m
* or use worst case (non fast-path above) with both types of algorithm.
*
! * - use -XX:+PrintGC to verify memory consumption of the benchmarks.
* (See "Checking Memory Consumption" below).
*
* Description:
*
* Fast-path algorithm for format(double...) call stack is very different of
--- 49,68 ----
* Note:
* - Vm options -Xms, -Xmx, -XX:NewSize must be set correctly for
* getting reliable numbers. Otherwise GC activity may corrupt results.
* As of jdk80b48 using "-Xms500m -Xmx500m -XX:NewSize=400m" covers
* all cases.
! * - Optionally using "-Xlog:gc" option provides information that
* helps checking any GC activity while benches are run.
*
* Vm Options:
* - Vm options to use (as of jdk80b48):
* fast-path case : -Xms128m -Xmx128m -XX:NewSize=100m
* non fast-path case: -Xms500m -Xmx500m -XX:NewSize=400m
* or use worst case (non fast-path above) with both types of algorithm.
*
! * - use -Xlog:gc to verify memory consumption of the benchmarks.
* (See "Checking Memory Consumption" below).
*
* Description:
*
* Fast-path algorithm for format(double...) call stack is very different of
*** 164,174 ****
*
* Running with "-verbose" arguments will provide clues of the memory consumed
* but is not enough, since any unexpected incremental GC may lower
* artificially the estimation of the memory consumption.
*
! * Options to set are -Xms, -Xmx, -XX:NewSize, plus -XX:+PrintGC to evaluate
* correctly the values of these options. When running "-verbose", varying
* numbers reported for memory consumption may indicate bad choices for these
* options.
*
* For jdk80b25, fast-path shows a consuption of ~60Mbs for 500000 iterations
--- 164,174 ----
*
* Running with "-verbose" arguments will provide clues of the memory consumed
* but is not enough, since any unexpected incremental GC may lower
* artificially the estimation of the memory consumption.
*
! * Options to set are -Xms, -Xmx, -XX:NewSize, plus -Xlog:gc to evaluate
* correctly the values of these options. When running "-verbose", varying
* numbers reported for memory consumption may indicate bad choices for these
* options.
*
* For jdk80b25, fast-path shows a consuption of ~60Mbs for 500000 iterations
*** 215,225 ****
"Note: \n" +
" - Vm options -Xms, -Xmx, -XX:NewSize must be set correctly for \n" +
" getting reliable numbers. Otherwise GC activity may corrupt results.\n" +
" As of jdk80b48 using \"-Xms500m -Xmx500m -XX:NewSize=400m\" covers \n" +
" all cases.\n" +
! " - Optionally using \"-XX:+printGC\" option provides information that \n" +
" helps checking any GC activity while benches are run.\n\n" +
"Look at the heading comments and description in source code for " +
"detailed information.\n");
}
--- 215,225 ----
"Note: \n" +
" - Vm options -Xms, -Xmx, -XX:NewSize must be set correctly for \n" +
" getting reliable numbers. Otherwise GC activity may corrupt results.\n" +
" As of jdk80b48 using \"-Xms500m -Xmx500m -XX:NewSize=400m\" covers \n" +
" all cases.\n" +
! " - Optionally using \"-Xlog:gc\" option provides information that \n" +
" helps checking any GC activity while benches are run.\n\n" +
"Look at the heading comments and description in source code for " +
"detailed information.\n");
}
< prev index next >