1LLVM-REDUCE(1)                       LLVM                       LLVM-REDUCE(1)
2
3
4

NAME

6       llvm-reduce - LLVM automatic testcase reducer.
7

SYNOPSIS

9       llvm-reduce [options] [input...]
10

DESCRIPTION

12       The  llvm-reduce tool project that can be used for reducing the size of
13       LLVM test cases.  It works by removing redundant  or  unnecessary  code
14       from  LLVM  test  cases  while still preserving their ability to detect
15       bugs.
16
17       If input is "-", llvm-reduce reads from standard input.  Otherwise,  it
18       will read from the specified filenames.
19
20       LLVM-Reduce  is  a  useful tool for reducing the size and complexity of
21       LLVM test cases, making it easier to identify and debug issues  in  the
22       LLVM compiler infrastructure.
23

GENERIC OPTIONS

25       --help Display available options (--help-hidden for more).
26
27       --abort-on-invalid-reduction
28              Abort if any reduction results in invalid IR
29
30       --in-place
31              WARNING:  This  option will replace your input file with the re‐
32              duced version!
33
34       --ir-passes=<string>
35              A textual description of  the  pass  pipeline,  same  as  what's
36              passed to opt -passes.
37
38       -j <uint>
39              Maximum  number of threads to use to process chunks. Set to 1 to
40              disable parallelism.
41
42       --max-pass-iterations=<int>
43              Maximum number of times to run the full set of delta passes (de‐
44              fault=5).
45
46       --mtriple=<string>
47              Set the target triple.
48
49       --preserve-debug-environment
50              Don't  disable features used for crash debugging (crash reports,
51              llvm-symbolizer and core dumps)
52
53       --print-delta-passes
54              Print list of delta passes,  passable  to  --delta-passes  as  a
55              comma separated liste.
56
57       --skip-delta-passes=<string>
58              Delta  passes  to  not run, separated by commas. By default, run
59              all delta passes.
60
61       --starting-granularity-level=<uint>
62              Number of times to divide chunks prior to first test.
63
64              Note : Granularity refers to the level of detail  at  which  the
65              reduction  process operates.  A lower granularity means that the
66              reduction process operates at a more coarse-grained level, while
67              a   higher   granularity  means  that  it  operates  at  a  more
68              fine-grained level.
69
70       --test=<string>
71              Name of the interesting-ness test to be run.
72
73       --test-arg=<string>
74              Arguments passed onto the interesting-ness test.
75
76       --verbose
77              Print extra debugging information.
78
79       --write-tmp-files-as-bitcode
80              Always write temporary files as bitcode instead of textual IR.
81
82       -x={ir|mir}
83              Input language as ir or mir.
84

EXIT STATUS

86       llvm-reduce returns 0 under normal operation.  It  returns  a  non-zero
87       exit code if there were any errors.
88

AUTHOR

90       Maintained by the LLVM Team (https://llvm.org/).
91

93       2003-2023, LLVM Project
94
95
96
97
9817                                2023-11-28                    LLVM-REDUCE(1)
Impressum