RFR (XS): 8007776: Test6852078.java timeouts
Vladimir Kozlov
vladimir.kozlov at oracle.com
Tue Feb 26 12:07:06 PST 2013
Fix code style:
< if ( i > 100 && System.currentTimeMillis() - start > 100000 )
---
> if ((i > 100) && ((System.currentTimeMillis() - start) > 100000))
Vladimir
On 2/26/13 11:43 AM, David Chase wrote:
> http://cr.openjdk.java.net/~drchase/8007776/webrev.01/
>
> Made changes, reran tests, rechecked early exit with extra iterations.
>
> On 2013-02-26, at 12:01 PM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:
>
>> David,
>>
>> Can you do it as separate check inside loop with 'break;' with a comment? I think it would be easier to read the code such way.
>>
>> Also you can remove /othervm since no flags are used?
>>
>> Thanks,
>> Vladimir
>>
>> On 2/26/13 6:24 AM, David Chase wrote:
>>>
>>> http://cr.openjdk.java.net/~drchase/8007776/webrev.00/
>>>
>>> Bug: Test6852078.java contained a 2000-test-iteration loop for tickling a sometimes-bug,
>>> that also sometimes timed out after as few as 130 iterations on some slower machines.
>>>
>>> Fix: Inserted an early-out on the test-loop -- if more than 100 seconds and more than
>>> 100 iterations have both passed, then exit is allowed.
>>> 2000 iterations are still the norm on faster machines (e.g., on a modern Intel laptop it
>>> takes 5 seconds to run).
>>>
>>> Testing:
>>> Jtreg, JPRT on that test.
>>> Hand-tested the early-out logic by bumping the iteration count to 2 million; it exited after
>>> a total test execution time of 100.5 seconds.
>>>
>
More information about the hotspot-compiler-dev
mailing list