the Determining the sublist of a list of candidate advisors that are applicable to the given class can be simply parallelized using the parallelStream to speed up the spring-context start-up process.
Comment From: pivotal-cla
@caitsithx Please sign the Contributor License Agreement!
Click here to manually synchronize the status of this Pull Request.
See the FAQ for frequently asked questions.
Comment From: pivotal-cla
@caitsithx Thank you for signing the Contributor License Agreement!
Comment From: snicoll
@caitsithx thanks for the suggestion but we don't intend to use parallel streams internally anywhere by design. If we parallelize, then at a higher level where we manage the granularity ourselves.
to speed up the spring-context start-up process.
Did you measure it? Given how much micro-switching this would involve at that level, we're very much doubtful this would actually make a difference. Going forward, please consider measuring and sharing a performance benchmarks as part of a a change like this.
Comment From: caitsithx
@caitsithx thanks for the suggestion but we don't intend to use parallel streams internally anywhere by design. If we parallelize, then at a higher level where we manage the granularity ourselves.
to speed up the spring-context start-up process.
Did you measure it? Given how much micro-switching this would involve at that level, we're very much doubtful this would actually make a difference. Going forward, please consider measuring and sharing a performance benchmarks as part of a a change like this.
Thank you @snicoll for the reviewing. 1) Got it. a fine-grained granularity could be more efficient, controllable and more tune-able. I will make some changes. 2) In fact I am working with a spring application with several thousands of beans, it could remove 1 mins during the startup by parallelizing the Determining of the candidate advisors. as you suggested, I will attach a performance test result.