Review request:7197210: java/lang/invoke/CallSiteTest.java failing on armsflt

Jiangli Zhou jiangli.zhou at oracle.com
Fri Nov 2 15:05:11 PDT 2012


Hi Vladimir,

Thanks for the review!

Jiangli

On 11/02/2012 03:03 PM, Vladimir Kozlov wrote:
> Looks good.
>
> We know that VerifyDependencies has significant effect on debug VM 
> performance.
>
> Thanks,
> Vladimir
>
> Jiangli Zhou wrote:
>> Redirecting the review request to core-libs-dev at openjdk.java.net mail 
>> list (second try) ...
>>
>> Here is the webrev based on the jdk8/tl/jdk repository:
>>
>> http://cr.openjdk.java.net/~jiangli/7197210/webrev.02/
>>
>> The '-XX:+IgnoreUnrecognizedVMOptions -XX:-VerifyDependencies' 
>> options are added to following tests to reduce workload. The 
>> -DRicochetTest.MAX_ARITY of RicochetTest is changed from 50 to10 to 
>> reduce execution time on slower device and binary. Timeout settings 
>> are also added to each test. The '-esa' option is added to 
>> BigArityTest and MethodHandlesTest to enable asserts.
>>
>>   java.lang.invoke.BigArityTest
>>   java.lang.invoke.MethodHandlesTest
>>   java.lang.invoke.CallSiteTest
>>   java.lang.invoke.RicochetTest
>>
>> Thanks,
>> Jiangli
>>
>> On 10/30/2012 11:38 AM, Jiangli Zhou wrote:
>>> Hi Chris,
>>>
>>> Here is the updated webrev with added 
>>> '-XX:+IgnoreUnrecognizedVMOptions -XX:-VerifyDependencies' options 
>>> and timeout setting for the following tests:
>>>
>>>   test.java.lang.invoke.RicochetTest
>>>   test.java.lang.invoke.BigArityTest
>>>   test.java.lang.invoke.MethodHandlesTest
>>>   test.java.lang.invoke.CallSiteTest
>>>
>>> http://cr.openjdk.java.net/~jiangli/7197210/webrev.01/
>>>
>>> Thanks,
>>>
>>> Jiangli
>>
>>



More information about the hotspot-dev mailing list