RFR(L): 8028468: Add inlining information into ciReplay
Christian Thalinger
christian.thalinger at oracle.com
Fri Jan 3 12:02:33 PST 2014
On Jan 3, 2014, at 11:58 AM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:
>
> On 1/3/14 11:42 AM, Christian Thalinger wrote:
>> Why are we using a different file for the inline data?
>
> I want to keep it smaller when we want only replay inlining. They have only one line with "compile" command and inline information. So they can't be used for normal compilation replay. They are created and used only with commands DumpInline and ReplayInline. Otherwise inlining info is included into replay_*.log file with other replay info.
So we can easily replay the inlining while running the application. That makes sense. Let me finish the review…
>
> Thanks,
> Vladimir
>
>>
>> On Dec 27, 2013, at 6:15 PM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:
>>
>>> http://cr.openjdk.java.net/~kvn/8028468/webrev/
>>>
>>> https://bugs.openjdk.java.net/browse/JDK-8028468
>>>
>>> Add inlining information into Compilation Replay data to correctly inline methods during recompilation.
>>> Add ability to use Replay data to replay compilation DURING normal execution of a program (in debug VM only because ciReplay is defined only in debug VM). Currently a program is not executed when we do recompilation.
>>> Allow dump and replay inlining for specified method during a program execution.
>>>
>>> Thanks,
>>> Vladimir
>>>
>>
More information about the hotspot-compiler-dev
mailing list