Pat, Thanks for adding the new test results. This idea for this implementation was Gábor's from the FLINK-3722 description.
Since you will be filing a FLIP I recommend including these benchmarks for consideration and discussion on the mailing list. In part because the PR is 4 months old and need of further review but also that there may be new ideas or questions as to the form of the new code. Greg On Tue, Feb 14, 2017 at 5:40 AM, Gábor Gévay <gga...@gmail.com> wrote: > Hello, > > Pat, the table in your email is somehow not visible in my gmail, but > it is visible here: > http://apache-flink-mailing-list-archive.1008284.n3. > nabble.com/FLINK-5734-Code-Generation-for-NormalizedKeySorter-tt15804. > html#a15936 > Maybe the problem is caused by the formatting. > > > FLINK-3722 > > approach seems to be the fastest one. > > OK, then I would suggest to do the implementation of the code > generation on top of the PR for FLINK-3722. I guess we can assume that > that PR will be merged sooner than the code generation, so there won't > be any serious merge conflicts this way. > > Best, > Gábor > > > > > 2017-02-14 11:16 GMT+01:00 pat.chormai <pat.chor...@gmail.com>: > > Hi [~greghogan] > > > > I have done the benchmark comparing between FLINK-3722 and our > approaches. > > As you can see at *Score * column which represents sorting time, > FLINK-3722 > > approach seems to be the fastest one. > > > > > > > > > > > > -- > > View this message in context: http://apache-flink-mailing- > list-archive.1008284.n3.nabble.com/FLINK-5734-Code-Generation-for- > NormalizedKeySorter-tp15804p15936.html > > Sent from the Apache Flink Mailing List archive. mailing list archive at > Nabble.com. >
