• CaffeineMark Instructions

• General Benchmarking Instructions

• Submitting Results

• Advanced Testing

CaffeineMark Instructions

To run the standard test group

1. Click on the Start Test button.
2. The test results will be displayed in the main test area as the test runs. A progress meter shows you the status of the benchmark.
3. The total testing time will be about 40 seconds.

To run the Local tests

1. Click on the Local... button.
2. Click on the button appropriate button for the test you wish to run.
3. The AGC test will take approximately 3 minutes to run.
4. The DC test will take just a few seconds.

General Benchmarking Instructions

Run the benchmark repeatedly until the result is stable. Our testing indicates that the CaffeineMark produces quite stable results, even on the first run. If your benchmark does not stabilize, it may be because of one of the following:

1. Your machine has insufficient memory to run the Java VM and Browser. This may cause your system to start swapping memory to disk.
2. Your graphics subsystem is batching graphics commands differently for each test run.
3. Your machine is a network server and is being accessed by other users during the test.
4. You are running an earlier variant of Java on the Macintosh. Early implementations of Mac Java have a time resolution of 1 second, as opposed to 1 millisecond (PC platforms typically have 20 millisecond resolution). This may cause the results to alternate between two different scores.

Summary

• Shutdown other unrelated tasks, e.g., mailreaders, spreadsheets, wordprocessors, ray tracers, etc.
• Run the test repeatedly until the result stabilizes.
• Avoid using other applications during the test. Keep your browser in the foreground and keep the mouse stationary.

Submitting Results

To submit local test results, you can send email to benchmark@webfayre.com with the following information:

1. Your Name and Company name.
2. Your system name, e.g., Micron Millenia P133 Plus
3. The Operating System you are using.
4. Your CPU type, speed, and L2 cache, e.g. 486DX4-100, 256K cache
5. The amount of RAM in your system
6. Very Important - the browser or Java appletviewer you are using and its version number.
7. Your graphics adapter, e.g., Diamond Stealth 64 PCI. Include your X-Windows version.
8. Your CaffeineMark
   • The scores on your individual tests. Impotrant!
   • The overall CaffeineMark for your system.
9. Comments or observations about the benchmark.

Advanced Testing

Also included in CaffeineMark 2.5 is the original String benchmark called STRING2. STRING2 is not used to calculate the CaffeineMark, but it is useful for isolating text processing performance. The difference between the two tests is that String does not allocate as much memory as String2, so it focusses much more on the string processing than on memory management. The standard String test is memory intensive.

To specify the test you wish to run, create an HTML file that looks like the following:
<APPLET CODE="CaffeineMark.class" WIDTH=500 HEIGHT=300>
<PARAM NAME="SPECIFIC" VALUE="STRING2">
<PARAM NAME="ITERATE" VALUE="5">
</APPLET>

The SPECIFIC parameter must have one of the following (case-sensitive) values:

• SIEVE
• LOOP
• LOGIC
• STRING
• STRING2
• FLOAT
• METHOD
• GRAPHICS
• IMAGE
• DIALOG