RFR(XS): JDK-8010941: MinJumpTableSize is set to 18, investigate if that's still optimal

[Niclas Adlertz]

Hi Aleksey, 

Thank you for your comments.

> At very least:
> * mean includes warmup time
As I see it, and please correct me if I'm wrong, warmup time should only be the first iteration of NUMBER_OF_TEST_EXECUTIONS (when doing 90000000 iterations (NUMBER_OF_ITERATIONS) and the on stack replacement triggers).
Ignoring the first test_time from the first iteration of NUMBER_OF_TEST_EXECUTIONS should be enough?

       if (i == 0) {
         System.out.println(" (ignored)");
       } else {
         overall_time += test_time;
         System.out.println("");
       }

> * store to $dead enables us to compute only the last iteration, and
> remove all others
What if I add to the dead value in each iteration of NUMBER_OF_ITERATIONS; 
 
       for (int j = 0; j < NUMBER_OF_ITERATIONS; ++j) {
          dead += multiply_by_power_of_ten(10.0, random_values[j]);
       } 
       long end_time = System.nanoTime();
       long test_time = end_time - start_time;

       System.out.print("[" + i + "]: " + test_time + " ignore " + dead);

and then also print dead in the end of each iteration of NUMBER_OF_TEST_EXECUTIONS? 
 
Please see the updated version of Test.java 
http://cr.openjdk.java.net/~adlertz/JDK-8010941/update/Test.java

Kind Regards,
Niclas Adlertz


On 21 maj 2013, at 14:57, Aleksey Shipilev wrote:

> On 05/21/2013 04:51 PM, Niclas Adlertz wrote:
>> Using the performance test http://cr.openjdk.java.net/~adlertz/JDK-8010941/Test.java  
> 
> Sorry, this benchmark is not trustworthy.
> 
> At very least:
> * mean includes warmup time
> * store to $dead enables us to compute only the last iteration, and
> remove all others
> 
> Please consider to re-validate your experiment using JMH (I can provide
> first-hand assistance with it):
>  http://openjdk.java.net/projects/code-tools/jmh/
> 
> -Aleksey.