FULL PRODUCT VERSION :
java version "1.6.0_17"
Java(TM) SE Runtime Environment (build 1.6.0_17-b04)
Java HotSpot(TM) 64-Bit Server VM (build 14.3-b01, mixed mode)

FULL OS VERSION :
Linux greycloud 2.6.31-17-generic #54-Ubuntu SMP Thu Dec 10 17:01:44 UTC 2009 x86_64 GNU/Linux

A DESCRIPTION OF THE PROBLEM :
See the test code below. Basically, sometimes simple integer multiplication and division fails.

THE PROBLEM WAS REPRODUCIBLE WITH -Xint FLAG: No

THE PROBLEM WAS REPRODUCIBLE WITH -server FLAG: Yes

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Run the JavaTest class given below (note that I've pasted two classes in there -- there's also a helper class). It only fails some of the time. Sometimes you have to run it several times (as many as 10) before you see the error.

EXPECTED VERSUS ACTUAL BEHAVIOR :
Expected:
64 * 64 / 8 == 512

Actual:
64 * 64 / 8 == 519
ERROR MESSAGES/STACK TRACES THAT OCCUR :
No error message, just wrong results.

REPRODUCIBILITY :
This bug can be reproduced often.

---------- BEGIN SOURCE ----------
package com.keith;

import java.io.IOException;

public class JavaTest
{
public static void main(String[] args) throws Exception
{
JavaTest macTest = new JavaTest();
for(long i = 0; i < 100000000l; i++)
{
macTest.doStuff();
}
}

private void doStuff() throws IOException {
JavaTestHelper dbbm = new JavaTestHelper();
dbbm.doMultiplicationAndDivision();
}

}


package com.keith;

import java.io.IOException;

public class JavaTestHelper
{
public int dimension = 4096;

public void doMultiplicationAndDivision() throws IOException
{
int length = (this.dimension + 63) / 64;
int arraySize = length * 64 / 8;
if(arraySize %2 == 1)
{
throw new RuntimeException("bad division!!");
}
}

}
---------- END SOURCE ----------

CUSTOMER SUBMITTED WORKAROUND :
If I do the division before the multiplication, I don't get the problem. That is, if I do:
  int arraySize = length * (64 / 8);
instead of:
  int arraySize = length * 64 / 8;
then the problem doesn't show up.
But I can't realistically track down and test every place in our application where we do multiplication and division. I basically have zero confidence in our application because of this.

Release Regression From : 6u13-rev
The above release value was the last known release where this
bug was not reproducible. Since then there has been a regression.

Release Regression From : 6u13-rev
The above release value was the last known release where this
bug was not reproducible. Since then there has been a regression.