Thanks for running the benchmark! I suspect your patch removed some of the "startup" portion of the cost that was previously associated with a call to layoutGlyphVector or LineBreakMeasurer--that's a very good step in the right direction.
-- Eirik From: Dmitry Batrak <[email protected]> Sent: Tuesday, March 12, 2019 4:36 AM To: Eirik Bakke <[email protected]> Cc: 2d-dev <[email protected]> Subject: Re: [OpenJDK 2D-Dev] [PATCH] 8220231: Cache HarfBuzz face object for same font's text layout calls > I wonder if this patch might also fix or improve > https://bugs.openjdk.java.net/browse/JDK-8202131 ? In case it's of any use, I > have a ready-made performance test for this at > https://gist.github.com/eirikbakke/81eec72221c9d7d8fc0b4afa9c54a163 . The proposed change improves the result of your performance test significantly, but it's still about two times slower as compared to 1.8.0_202 results on my machine. Best regards, Dmitry Batrak On Fri, Mar 8, 2019 at 10:54 PM Eirik Bakke <[email protected]<mailto:[email protected]>> wrote: A random Java developer here passing by... great to see work on HarfBuzz performance! Thanks for your efforts to push JetBrains JDK improvements upstream. I wonder if this patch might also fix or improve https://bugs.openjdk.java.net/browse/JDK-8202131 ? In case it's of any use, I have a ready-made performance test for this at https://gist.github.com/eirikbakke/81eec72221c9d7d8fc0b4afa9c54a163 . -- Eirik From: 2d-dev <[email protected]<mailto:[email protected]>> On Behalf Of Dmitry Batrak Sent: Wednesday, March 6, 2019 7:15 AM To: 2d-dev <[email protected]<mailto:[email protected]>> Subject: [OpenJDK 2D-Dev] [PATCH] 8220231: Cache HarfBuzz face object for same font's text layout calls Hello, I'd like to submit a patch for JDK-8220231. I'm not a Committer, so I'll need someone to sponsor this change. The proposed approach is used without known issues in OpenJDK-based JetBrains Runtime for almost three years now. I've mentioned it previously on this mailing list (https://mail.openjdk.java.net/pipermail/2d-dev/2017-August/008497.html). The change has been refactored as compared to the version mentioned above (the logic has been moved to SunLayoutEngine), and includes the removal of font tables caching (JDK-8186317). The latter, I believe, becomes redundant with this fix. Issue: https://bugs.openjdk.java.net/browse/JDK-8220231 Webrev: http://cr.openjdk.java.net/~dbatrak/8220231/webrev.00/ Best regards, Dmitry Batrak
