[OpenJDK 2D-Dev] [9] Review Request: 8140266 Performance loss between jdk8 and jdk9 on Maskfill

Sergey Bylokhov sergey.bylokhov at oracle.com
Mon Dec 26 15:06:40 UTC 2016


If there are no objections I’ll push this version:
http://cr.openjdk.java.net/~serb/8140266/webrev.01/make/lib/Awt2dLibraries.gmk.sdiff.html <http://cr.openjdk.java.net/~serb/8140266/webrev.01/make/lib/Awt2dLibraries.gmk.sdiff.html>

> 
> Build change looks ok, but I too think a comment explaining why is a good idea.
> 
> /Erik
> 
> 
> On 2016-12-23 15:49, Philip Race wrote:
>> Looks OK to me although it is a very cryptic option so maybe include a comment
>> that it is for performance with some versions of gcc ...
>> 
>> -phil.
>> 
>> On 12/23/16, 3:15 AM, Sergey Bylokhov wrote:
>>> Hello.
>>> Please review the small fix for jdk9.
>>> The change add an additional gcc option to the libawt lib to improve some graphics operations: 20%.
>>> 
>>> This option is a part of -03.
>>>> |-fgcse-after-reload|
>>>>    When -fgcse-after-reload is enabled, a redundant load elimination
>>>>    pass is performed after reload. The purpose of this pass is to
>>>>    clean up redundant spilling. 
>>> https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html
>>> 
>>> Note that if -03 will be enabled directly via OPTIMIZATION flag then the performance will not be improved.
>>> 
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8140266
>>> Webrev can be found at: http://cr.openjdk.java.net/~serb/8140266/webrev.00
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/2d-dev/attachments/20161226/7a96543d/attachment.html>


More information about the 2d-dev mailing list