MessageCracker.xsl optimization using HashMap

[busy-spin]

Fixes #577

[busy-spin]

Following is the JMH benchmarking results.

Before:

benchmarkCrackingGenerated - use the current MessageCracker.xls generated class
benchmarkCrackingWithHashMap use custom MessageCracker class which use HashMaps to resolve methods

Benchmark Mode Cnt Score Error Units
MessageCrackerBenchmark.benchmarkCrackingGenerated thrpt 5 10923194.106 ± 2861197.934 ops/s
MessageCrackerBenchmark.benchmarkCrackingWithHashMap thrpt 5 69723897.515 ± 3242125.694 ops/s

After:

benchmarkCrackingGenerated - use the new MessageCracker.xls generated class
benchmarkCrackingWithHashMap use custom MessageCracker class which use HashMaps to resolve methods

Benchmark Mode Cnt Score Error Units
MessageCrackerBenchmark.benchmarkCrackingGenerated thrpt 5 71897498.754 ± 19147681.365 ops/s
MessageCrackerBenchmark.benchmarkCrackingWithHashMap thrpt 5 72165054.010 ± 4905711.116 ops/s

To Reproduce
Run my JMH benchmark code in my forked repo for QFJ

Before

Before Code Benchmark

After

After Code Benchmark

[chrjohn]

@chrjohn need to check if new code generator for FIX Latest can benefit from this.

[mrFloony]

hi, I implemented my version of MessageCracker based on HashMap, and he showed a deterioration in processing speed (~30 ms difference compared to If-Else on one message). The replacement with a Switch turned out to be a little faster (a gain of ~15 ms with the same parameters). Tested on a pre-battle stand (with slightly reduced parameters of the virtual machine and allocated memory), as a result, the implementation on the Switch went to the product stand (three weeks - normal flight)
Therefore, I would rather replace the basic MessageCracker with an implementation from Switch

[chrjohn]

@mrFloony thanks for your comment. Are you maybe able to create a benchmark as linked in #578 (comment) or as used in the quickfixj-perf-test module #582

[busy-spin]

hi @mrFloony

Can let me know more details about your test. 30ms or 15ms for cracking message seems quiet high. Even the if-else approach is much faster than that. Do you have similar JMH test for this.

[busy-spin]

@chrjohn need to check if new code generator for FIX Latest can benefit from this.

Can let me know more details about this, I should be able to update my test. And also add comparison with switch statement vs hash map also.

[chrjohn]

@chrjohn need to check if new code generator for FIX Latest can benefit from this.

Can let me know more details about this, I should be able to update my test. And also add comparison with switch statement vs hash map also.

Actually this is in a different project which generates the classes from orchestrations: https://github.com/quickfix-j/quickfixj-orchestra/blob/fac2493c93f79e21ffcb3cb67cc25459dc2f467d/quickfixj-from-fix-orchestra-repository/quickfixj-from-fix-orchestra-generator/src/main/java/org/quickfixj/orchestra/CodeGeneratorJ.java#L577