[OpenJDK 2D-Dev] request for review: 9006475: xrender: improve performance of small fillRect operations

Phil Race philip.race at oracle.com
Wed Sep 4 17:44:31 UTC 2013

Looks fine to me too.
Clemsns: All the 9XXXXXX numbers are incident numbers. jcheck will reject
them as bug numbers.


On 9/4/2013 5:06 AM, Andrew Brygin wrote:
> Hello Clemens,
>  the change looks fine to me.
>  BTW, I have converted the report 9006475 to a jdk bug:
>  http://bugs.sun.com/view_bug.do?bug_id=8024261
>  It should be visible in the bugs database shortly.
> Thanks,
> Andrew
> On 9/4/2013 3:36 PM, Clemens Eisserer wrote:
>> Hi,
>> Please review my patch for 9006475 at: 
>> http://cr.openjdk.java.net/~ceisserer/9006475/webrev.00/ 
>> <http://cr.openjdk.java.net/%7Eceisserer/9006475/webrev.00/>
>> Before this patch we were calling the native method 
>> XRenderRectanglesNative() even in the case that there was only single 
>> rectangle to be rendered (which is the case for Graphics.fillRect). 
>> As XRenderRectanglesNative() uses GetPrimitiveArrayCritical, this 
>> caused high overhead for small fillRect() oerations.
>> With this patch we call renderRectangle instead, which passes the 
>> reactangle's coordinates down using JNI function parameters. On my 
>> private Laptop this improved throughput of 1x1 fillRect calls by a 
>> factor of 3.
>> Thanks, Clemens

More information about the 2d-dev mailing list