1LLVM-OPT-REPORT(1)                   LLVM                   LLVM-OPT-REPORT(1)
2
3
4

NAME

6       llvm-opt-report - generate optimization report from YAML
7

SYNOPSIS

9       llvm-opt-report [options] [input]
10

DESCRIPTION

12       llvm-opt-report  is a tool to generate an optimization report from YAML
13       optimization record files.
14
15       You need to create an input YAML optimization record file  before  run‐
16       ning llvm-opt-report.
17
18       It  provides information on the execution time, memory usage, and other
19       details of each optimization pass.
20
21          $ clang -c foo.c -o foo.o -O3 -fsave-optimization-record
22
23       Then, you create a report using the llvm-opt-report  command  with  the
24       YAML optimization record file foo.opt.yaml as input.
25
26          $ llvm-opt-report foo.opt.yaml -o foo.lst
27
28       foo.lst is the generated optimization report.
29
30          < foo.c
31           1          | void bar();
32           2          | void foo() { bar(); }
33           3          |
34           4          | void Test(int *res, int *c, int *d, int *p, int n) {
35           5          |   int i;
36           6          |
37           7          | #pragma clang loop vectorize(assume_safety)
38           8     V4,1 |   for (i = 0; i < 1600; i++) {
39           9          |     res[i] = (p[i] == 0) ? res[i] : res[i] + d[i];
40          10          |   }
41          11          |
42          12  U16     |   for (i = 0; i < 16; i++) {
43          13          |     res[i] = (p[i] == 0) ? res[i] : res[i] + d[i];
44          14          |   }
45          15          |
46          16 I        |   foo();
47          17          |
48          18          |   foo(); bar(); foo();
49             I        |   ^
50             I        |                 ^
51          19          | }
52          20          |
53
54       Symbols  printed  on the left side of the program indicate what kind of
55       optimization was performed.  The meanings of the symbols  are  as  fol‐
56       lows:
57
58       • I: The function is inlined.
59
60       • U:  The  loop  is unrolled. The following number indicates the unroll
61         factor.
62
63       • V: The loop is vectorized. The following numbers indicate the  vector
64         length and the interleave factor.
65
66       NOTE:
67          If  a  specific line of code is output twice, it means that the same
68          optimization pass was applied to that line of code  twice,  and  the
69          pass was able to further optimize the code on the second iteration.
70

OPTIONS

72       If  input is "-" or omitted, llvm-opt-report reads from standard input.
73       Otherwise, it will read from the specified filename.
74
75       If the -o option is omitted, then llvm-opt-report will send its  output
76       to  standard  output.   If the -o option specifies "-", then the output
77       will also be sent to standard output.
78
79       --help Display available options.
80
81       --version
82              Display the version of this program.
83
84       --format=<string>
85              The format of the optimization record file.  The Argument is one
86              of the following:
87
88              • yaml
89
90              • yaml-strtab
91
92              • bitstream
93
94       --no-demangle
95              Do not demangle function names.
96
97       -o=<string>
98              Output file.
99
100       -r=<string>
101              Root for relative input paths.
102
103       -s     Do not include vectorization factors, etc.
104

EXIT STATUS

106       llvm-opt-report  returns  0  on success. Otherwise, an error message is
107       printed to standard error, and the tool returns 1.
108

AUTHOR

110       Maintained by the LLVM Team (https://llvm.org/).
111

113       2003-2023, LLVM Project
114
115
116
117
11817                                2023-11-28                LLVM-OPT-REPORT(1)
Impressum