"42 < $a < 666" should be faster than "42 < $a && $a < 666", but is 2x as slow

[lizmat]

The Problem

"42 < $a < 666" should be faster than "42 < $a && $a < 666", but is 2x as slow

Expected Behavior

42 < $a < 666 should be faster

Actual Behavior

42 < $a < 666 is twice as slow

Steps to Reproduce

m: my $a = 42; for ^1000000 { my $b = 42 < $a < 666 }; say now - INIT now
<+camelia> rakudo-moar 7a4743b: OUTPUT: «0.54265194␤»
m: my $a = 42; for ^1000000 { my $b = 42 < $a && $a < 666 }; say now - INIT now
<+camelia> rakudo-moar 7a4743b: OUTPUT: «0.2490039␤»

Environment

• Operating system:

Darwin 16.7.0 Darwin Kernel Version 16.7.0: Thu Jun 15 17:36:27 PDT 2017; root:xnu-3789.70.16~2/RELEASE_X86_64 x86_64

• Compiler version (perl6 -v):

This is Rakudo version 2017.12-111-g8023d21 built on MoarVM version 2017.12-20-g5710340
implementing Perl 6.c.

[zoffixznet]

This is now fixed by b77d875 and accompanying Raku/nqp@4145205

For OP code, the perf is now about the same:

$ ./perl6 -e 'my $a = 42; for ^1000000 { my $b = 42 < $a && $a < 666 }; say now - INIT now'
0.2635152
$ ./perl6 -e 'my $a = 42; for ^1000000 { my $b = 42 < $a < 666 }; say now - INIT now'                                                                                                                           
0.2624662

And with longer chains, the chained version has noticeable perf benefits:

$ ./perl6 -e 'my $a = 42; for ^10_000_000 { my $b = 42 < $a && $a < 666 <= $a && $a <= 666 }; say now - INIT now'
2.6051729
$ ./perl6 -e 'my $a = 42; for ^10_000_000 { my $b = 42 < $a < 666 <= $a <= 666 }; say now - INIT now'
2.5022431

I guess I'm gonna close this without a test, since we don't yet have a solid perf testing framework...