http://bugzilla.novell.com/show_bug.cgi?id=524595
User [email protected] added comment http://bugzilla.novell.com/show_bug.cgi?id=524595#c4 --- Comment #4 from Zoltan Varga <[email protected]> 2009-07-24 20:33:24 MDT --- The last result shows only a 2x difference on amd64. Which could be expected since the slow version does more work inside the loop. The jit there generates the following code for the slow code: <BB>:5 20: 4d 8b ef mov %r15,%r13 23: 45 03 ee add %r14d,%r13d 26: 4d 8b fd mov %r13,%r15 <BB>:4 29: b8 90 b1 fd 01 mov $0x1fdb190,%eax 2e: 48 63 00 movslq (%rax),%rax 31: 44 3b f8 cmp %eax,%r15d 34: 7c ea jl 20 <nClass_Slow+0x20> and for the fast code: <BB>:5 18: 45 03 fe add %r14d,%r15d <BB>:4 1b: b8 90 01 51 02 mov $0x2510190,%eax 20: 48 63 00 movslq (%rax),%rax 23: 44 3b f8 cmp %eax,%r15d 26: 7c f0 jl 18 <nClass_Fast+0x18> -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug. _______________________________________________ mono-bugs maillist - [email protected] http://lists.ximian.com/mailman/listinfo/mono-bugs
