JFCMark Publishing Rules

We do not want to impose excessively strict rules, like those of SPEC, for reporting JFCMark results. You may use any ahead-of-time, "closed world", and feedback driven optimizations, pre-compilation and other techniques resulting in higher performance of JFCMark. You may vary the default settings of execution parameters as described in the "Configuring benchmarks" section of the main JFCMark readme file.

If you are a Java vendor or author of a JVM or a Java optimization tool and wish to publish your JFCMark test results, make sure to include the following information:

1. Identify Excelsior, LLC and the name and version number of the benchmark suite used. For instance:

   Excelsior, LLC's JFCMark v1.0

2. Specify the exact test name, e.g. InternalFrameDemo_drag and commands used to run it, for instance,

   runHS_drag.bat 10

   ("problem size" is 10, and "ratio" is not used by the test)

3. Mention whether optimizations unsafe for Java dynamic classloading were or were not applied, for instance:

   IBM JAX, an Application Extractor for Java performs a whole-program analysis unsafe w.r.t. loading new classes unknown at optimization time

4. Document the hardware configuration of the system under test. For example:

   CPU: AMD Athlon(tm) at 1,400 MHz with 256KB L2 cache
   RAM: 512 MB DDR
   Video: Asus V7100 Pro (GeForce2 MX 400) at 1024x768x65536c

5. Identify the operating system, version, and build used to run the tests, and patches or service packs installed. For example:

   Microsoft Windows NT Workstation Version 4.0 (Build 1381: Service Pack 6)

6. Identify the JVM used to run the tests, including the information printed by the "java -version" command, for instance:

   Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.1)
   Classic VM (build 1.3.1, J2RE 1.3.1 IBM Windows 32 build cn131-20020403 (JIT enabled: jitc))