Yui NARUSE wrote:
You know, this is bug tracker, not socks under the Christmas tree.
Moreover the benchmark you show is micro benchmark.
If we improve Ruby for such micro benchmark, the real performance of Ruby may not be improved.
Because such optimizations sometimes has trade offs.
We are trying to make more suitable benchmarks which reflect real world performance,
and improve Ruby for real world applications.
I did not want anything to offend. Sorry.
I said firstly: "Not BUG but wish"
May you point me to "real world performance", "real world applications" benchmarks? Source code for testing.
Many thanks!
And further...
Each complex ("real world") algorithm is a totality (set) of the many micro (simple) algorithms.
Disappointing that php is faster.
Sorry. I apologize again.
Ruby 2.3 is faster then 2.2.4!
test 1
Ruby 2.3.0:
user system total real
29000000
0.620000 0.000000 0.620000 ( 0.620483)
118000000
1.190000 0.000000 1.190000 ( 1.192848)
267000000
1.830000 0.000000 1.830000 ( 1.828014)
Ruby 2.2.4:
user system total real
29000000
0.810000 0.000000 0.810000 ( 0.811010)
118000000
1.640000 0.000000 1.640000 ( 1.641845)
267000000
2.480000 0.000000 2.480000 ( 2.476726)
test 2
Ruby 2.3.0:
user system total real
832040
0.140000 0.000000 0.140000 ( 0.140042)
1346269
0.180000 0.000000 0.180000 ( 0.178167)
2178309
0.290000 0.000000 0.290000 ( 0.288229)
3524578
0.460000 0.000000 0.460000 ( 0.464329)
5702887
0.760000 0.000000 0.760000 ( 0.756518)
Ruby 2.2.4:
user system total real
832040
0.170000 0.000000 0.170000 ( 0.165969)
1346269
0.210000 0.000000 0.210000 ( 0.214349)
2178309
0.350000 0.000000 0.350000 ( 0.346734)
3524578
0.560000 0.000000 0.560000 ( 0.562673)
5702887
0.910000 0.000000 0.910000 ( 0.904228)