|
Boost Users : |
From: Erik Thiele (erik.thiele_at_[hidden])
Date: 2019-05-24 13:41:21
I assume that time_duration is implemented as an integer. Some values
of this integer are then reserved for special values like
not_a_date_time.
this can be seen here:
https://github.com/boostorg/date_time/blob/develop/include/boost/date_time/int_adapter.hpp
this should have the consequence that when studying the assembler code
for comparing two time_durations, all that happens is the assembler
code of "less than compare" for integer type. if additional comparisons
and branchings take place, then I assume the code is not meeting the
design criteria of being optimized.
Well... this is what I did, copy paste this code to justcompare.cpp:
#include <boost/date_time/posix_time/ptime.hpp>
using namespace boost::posix_time;
bool compare_time_duration(time_duration a, time_duration b)
{
return a < b;
}
bool compare_int(int a, int b)
{
return a < b;
}
EOF
g++ -Wall -O2 -c justcompare.cpp
objdump -S --demangle justcompare.o
justcompare.o: file format elf64-x86-64
Disassembly of section .text:
0000000000000000
<compare_time_duration(boost::posix_time::time_duration, boost::posix_time::time_duration)>:
0: 48 8b 17 mov (%rdi),%rdx
3: 48 b9 ff ff ff ff ff movabs $0x7fffffffffffffff,%rcx
a: ff ff 7f
d: 48 8d 04 0a lea (%rdx,%rcx,1),%rax
11: 48 83 f8 fd cmp $0xfffffffffffffffd,%rax
15: 77 29 ja 40
17: 48 bf fe ff ff ff ff movabs $0x7ffffffffffffffe,%rdi
1e: ff ff 7f
21: 31 c0 xor %eax,%eax
23: 48 39 fa cmp %rdi,%rdx
26: 74 60 je 88
28: 48 8b 36 mov (%rsi),%rsi
2b: 48 01 f1 add %rsi,%rcx
2e: 48 83 f9 fd cmp $0xfffffffffffffffd,%rcx
32: 77 20 ja 54
34: 48 39 fe cmp %rdi,%rsi
37: 74 0a je 43
39: 48 39 f2 cmp %rsi,%rdx
3c: 0f 9c c0 setl %al
3f: c3 retq
40: 48 8b 36 mov (%rsi),%rsi
43: 48 b9 fe ff ff ff ff movabs $0x7ffffffffffffffe,%rcx
4a: ff ff 7f
4d: 31 c0 xor %eax,%eax
4f: 48 39 ce cmp %rcx,%rsi
52: 74 3c je 90
54: 48 bf 00 00 00 00 00 movabs $0x8000000000000000,%rdi
5b: 00 00 80
5e: 48 39 fa cmp %rdi,%rdx
61: 74 35 je 98
63: 48 b9 ff ff ff ff ff movabs $0x7fffffffffffffff,%rcx
6a: ff ff 7f
6d: 48 39 ce cmp %rcx,%rsi
70: 0f 94 c0 sete %al
73: 48 39 ca cmp %rcx,%rdx
76: 41 0f 95 c0 setne %r8b
7a: 44 20 c0 and %r8b,%al
7d: 74 21 je a0
7f: f3 c3 repz retq
81: 0f 1f 80 00 00 00 00 nopl 0x0(%rax)
88: f3 c3 repz retq
8a: 66 0f 1f 44 00 00 nopw 0x0(%rax,%rax,1)
90: f3 c3 repz retq
92: 66 0f 1f 44 00 00 nopw 0x0(%rax,%rax,1)
98: 48 39 d6 cmp %rdx,%rsi
9b: 0f 95 c0 setne %al
9e: c3 retq
9f: 90 nop
a0: 48 39 ca cmp %rcx,%rdx
a3: 74 da je 7f
a5: 48 39 fe cmp %rdi,%rsi
a8: 75 8f jne 39
aa: f3 c3 repz retq
ac: 0f 1f 40 00 nopl 0x0(%rax)
00000000000000b0 <compare_int(int, int)>:
b0: 39 f7 cmp %esi,%edi
b2: 0f 9c c0 setl %al
b5: c3 retq
(i removed the verbose jump target strings from the output to make the
email readable)
see the difference between compare_time_duration and compare_int...
I have a bad feeling in my stomach... sorry, something is going very
wrong here :-)
cya
erik
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net