RFR: JDK-8043674 : Update compiler/intrinsic/bmi tests to run it on all platforms.

Vladimir Kozlov vladimir.kozlov at oracle.com
Mon Sep 15 16:38:59 UTC 2014



On 9/15/14 6:42 AM, Dmitrij Pochepko wrote:
> Now with correct subject.
> Thanks,
> Dmitrij
>> The change looks good.
>> Thanks,
>> Filipp.
>> On 09/15/2014 04:31 PM, Dmitrij Pochepko wrote:
>>> Hi,
>>> Filipp have put it:
>>> http://cr.openjdk.java.net/~fzhinkin/dpochepk/8043674/webrev.00/
>>> Thanks,
>>> Dmitrij
>>>> Email subject should follow review request style.
>>>> Changes for review should be sent as webrev on cr.openjdk server. Ask Filipp or Igor to put it there for you. They
>>>> also can be your sponsors. Filipp is these tests author and should review changes too.
>>>> I am fine with running tests on all platforms.
>>>> Thanks,
>>>> Vladimir
>>>> On 9/12/14 1:38 PM, dmitrij pochepko wrote:
>>>>> Hi hotspot compiler team,
>>>>> I'm a member of java sqe group, and currently i'm proposing a small fix for hotspot tests.
>>>>> A respective Jira issue is https://bugs.openjdk.java.net/browse/JDK-8043674 - Update compiler/intrinsic/bmi tests
>>>>> to run
>>>>> it on all platforms.
>>>>> A diff is attached. A bmi-related tests compare vm work using Xint and Xcomp modes(assuming Xcomp uses bmi
>>>>> instructions). In case of running on another platforms it's just usual comparison of Xint and Xcomp modes, which won't
>>>>> test bmi, however, some platforms might have some bmi-like optimizations in future(for example, a powerpc platform
>>>>> have
>>>>> the some of such instructions), so, it's only natural to allow bmi tests to be run anywhere.
>>>>> What i am asking is some reviewers for this diff and(in case review is successfull) commiter to do the push(since i
>>>>> have
>>>>> no commiter status).
>>>>> Thanks,
>>>>> Dmitrij

More information about the hotspot-compiler-dev mailing list