On Fri, 8 Aug 2025 15:31:18 GMT, Francesco Andreuzzi <d...@openjdk.org> wrote:
>> You need to modify the comment here, too. > > Right, I'll wait just a bit so the discussion on how to approach the problem > stabilizes :) This number is based on my original experimentation. If I tried to lower the bar by lowering the number, the PCH list grew too much and it made for a worse performance. And contrary, if I raised the number, fewer files where included which made the PCH quicker to process but less helpful. That number was the optimum I found. It seems from https://bugs.openjdk.org/browse/JDK-8365053 that this is still around the optimum. However, rather than just mentioning the number in the comment, the rationale could be specified, like `the optimum number of includes`. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/26681#discussion_r2265991618
