gitgabrio commented on PR #5621: URL: https://github.com/apache/incubator-kie-drools/pull/5621#issuecomment-1859727238
HI @tarilabs , thanks for help! Reading the raw data, I see that for `evaluateCompiledButInterpretedExpression` there is a loss of performance. But for `evaluateCompiledJavaExpressionBenchmark` there is a mix of result: in some cases a loss, in some cases a gain. I'll look to improve that, and it would be extemely helpful if you could share some suggestion on what to benchmark exactly, namely: 1. I picked some infix expression from the code base, but I'm not sure if those are actually representative of the business domain 2. between the three different codepath executions I see in the benchmark, I'm not sure which is the most important, i.e. which one should I focus on the mostly Could you please share them on the [benchmark PR](https://github.com/apache/incubator-kie-benchmarks/pull/279/) ? @mariofusco @tarilabs @baldimir That PR basically does two modifications 1. move code to specific packages/classes 2. remove the `if-driven` logic with a sort of "class-pattern-matching" based on the `ClassIdentifierTuple` Compared with the "if" choice, I'm pretty sure that the second point is more expensive, but removing that "if" was one of the goal: do you have some idea/suggestion for a more efficient implementation ? The requirement is that for any given pair of class types, a specific method should be invoked. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
