SampleProfileLoader pass. Initial setup.
This adds a new scalar pass that reads a file with samples generated
by 'perf' during runtime. The samples read from the profile are
incorporated and emmited as IR metadata reflecting that profile.
The profile file is assumed to have been generated by an external
profile source. The profile information is converted into IR metadata,
which is later used by the analysis routines to estimate block
frequencies, edge weights and other related data.
External profile information files have no fixed format, each profiler
is free to define its own. This includes both the on-disk representation
of the profile and the kind of profile information stored in the file.
A common kind of profile is based on sampling (e.g., perf), which
essentially counts how many times each line of the program has been
executed during the run.
The SampleProfileLoader pass is organized as a scalar transformation.
On startup, it reads the file given in -sample-profile-file to
determine what kind of profile it contains. This file is assumed to
contain profile information for the whole application. The profile
data in the file is read and incorporated into the internal state of
the corresponding profiler.
To facilitate testing, I've organized the profilers to support two file
formats: text and native. The native format is whatever on-disk
representation the profiler wants to support, I think this will mostly
be bitcode files, but it could be anything the profiler wants to
support. To do this, every profiler must implement the
SampleProfile::loadNative() function.
The text format is mostly meant for debugging. Records are separated by
newlines, but each profiler is free to interpret records as it sees fit.
Profilers must implement the SampleProfile::loadText() function.
Finally, the pass will call SampleProfile::emitAnnotations() for each
function in the current translation unit. This function needs to
translate the loaded profile into IR metadata, which the analyzer will
later be able to use.
This patch implements the first steps towards the above design. I've
implemented a sample-based flat profiler. The format of the profile is
fairly simplistic. Each sampled function contains a list of relative
line locations (from the start of the function) together with a count
representing how many samples were collected at that line during
execution. I generate this profile using perf and a separate converter
tool.
Currently, I have only implemented a text format for these profiles. I
am interested in initial feedback to the whole approach before I send
the other parts of the implementation for review.
This patch implements:
- The SampleProfileLoader pass.
- The base ExternalProfile class with the core interface.
- A SampleProfile sub-class using the above interface. The profiler
generates branch weight metadata on every branch instructions that
matches the profiles.
- A text loader class to assist the implementation of
SampleProfile::loadText().
- Basic unit tests for the pass.
Additionally, the patch uses profile information to compute branch
weights based on instruction samples.
This patch converts instruction samples into branch weights. It
does a fairly simplistic conversion:
Given a multi-way branch instruction, it calculates the weight of
each branch based on the maximum sample count gathered from each
target basic block.
Note that this assignment of branch weights is somewhat lossy and can be
misleading. If a basic block has more than one incoming branch, all the
incoming branches will get the same weight. In reality, it may be that
only one of them is the most heavily taken branch.
I will adjust this assignment in subsequent patches.
llvm-svn: 194566
2013-11-13 20:22:21 +08:00
|
|
|
; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/branch.prof | opt -analyze -branch-prob | FileCheck %s
|
|
|
|
|
|
|
|
|
|
; Original C++ code for this test case:
|
|
|
|
|
;
|
|
|
|
|
; #include <stdio.h>
|
|
|
|
|
; #include <stdlib.h>
|
|
|
|
|
;
|
|
|
|
|
; int main(int argc, char *argv[]) {
|
|
|
|
|
; if (argc < 2)
|
|
|
|
|
; return 1;
|
|
|
|
|
; double result;
|
|
|
|
|
; int limit = atoi(argv[1]);
|
|
|
|
|
; if (limit > 100) {
|
|
|
|
|
; double s = 23.041968;
|
|
|
|
|
; for (int u = 0; u < limit; u++) {
|
|
|
|
|
; double x = s;
|
|
|
|
|
; s = x + 3.049 + (double)u;
|
|
|
|
|
; s -= s + 3.94 / x * 0.32;
|
|
|
|
|
; }
|
|
|
|
|
; result = s;
|
|
|
|
|
; } else {
|
|
|
|
|
; result = 0;
|
|
|
|
|
; }
|
|
|
|
|
; printf("result is %lf\n", result);
|
|
|
|
|
; return 0;
|
|
|
|
|
; }
|
|
|
|
|
|
|
|
|
|
@.str = private unnamed_addr constant [15 x i8] c"result is %lf\0A\00", align 1
|
|
|
|
|
|
|
|
|
|
; Function Attrs: nounwind uwtable
|
|
|
|
|
define i32 @main(i32 %argc, i8** nocapture readonly %argv) #0 {
|
|
|
|
|
; CHECK: Printing analysis 'Branch Probability Analysis' for function 'main':
|
|
|
|
|
|
|
|
|
|
entry:
|
|
|
|
|
tail call void @llvm.dbg.value(metadata !{i32 %argc}, i64 0, metadata !13), !dbg !27
|
|
|
|
|
tail call void @llvm.dbg.value(metadata !{i8** %argv}, i64 0, metadata !14), !dbg !27
|
|
|
|
|
%cmp = icmp slt i32 %argc, 2, !dbg !28
|
|
|
|
|
br i1 %cmp, label %return, label %if.end, !dbg !28
|
|
|
|
|
; CHECK: edge entry -> return probability is 1 / 2 = 50%
|
|
|
|
|
; CHECK: edge entry -> if.end probability is 1 / 2 = 50%
|
|
|
|
|
|
|
|
|
|
if.end: ; preds = %entry
|
|
|
|
|
%arrayidx = getelementptr inbounds i8** %argv, i64 1, !dbg !30
|
|
|
|
|
%0 = load i8** %arrayidx, align 8, !dbg !30, !tbaa !31
|
|
|
|
|
%call = tail call i32 @atoi(i8* %0) #4, !dbg !30
|
|
|
|
|
tail call void @llvm.dbg.value(metadata !{i32 %call}, i64 0, metadata !17), !dbg !30
|
|
|
|
|
%cmp1 = icmp sgt i32 %call, 100, !dbg !35
|
|
|
|
|
br i1 %cmp1, label %for.body, label %if.end6, !dbg !35
|
|
|
|
|
; CHECK: edge if.end -> for.body probability is 2243 / 2244 = 99.9554% [HOT edge]
|
|
|
|
|
; CHECK: edge if.end -> if.end6 probability is 1 / 2244 = 0.0445633%
|
|
|
|
|
|
|
|
|
|
for.body: ; preds = %if.end, %for.body
|
|
|
|
|
%u.016 = phi i32 [ %inc, %for.body ], [ 0, %if.end ]
|
|
|
|
|
%s.015 = phi double [ %sub, %for.body ], [ 0x40370ABE6A337A81, %if.end ]
|
|
|
|
|
%add = fadd double %s.015, 3.049000e+00, !dbg !36
|
|
|
|
|
%conv = sitofp i32 %u.016 to double, !dbg !36
|
|
|
|
|
%add4 = fadd double %add, %conv, !dbg !36
|
|
|
|
|
tail call void @llvm.dbg.value(metadata !{double %add4}, i64 0, metadata !18), !dbg !36
|
|
|
|
|
%div = fdiv double 3.940000e+00, %s.015, !dbg !37
|
|
|
|
|
%mul = fmul double %div, 3.200000e-01, !dbg !37
|
|
|
|
|
%add5 = fadd double %add4, %mul, !dbg !37
|
|
|
|
|
%sub = fsub double %add4, %add5, !dbg !37
|
|
|
|
|
tail call void @llvm.dbg.value(metadata !{double %sub}, i64 0, metadata !18), !dbg !37
|
|
|
|
|
%inc = add nsw i32 %u.016, 1, !dbg !38
|
|
|
|
|
tail call void @llvm.dbg.value(metadata !{i32 %inc}, i64 0, metadata !21), !dbg !38
|
|
|
|
|
%exitcond = icmp eq i32 %inc, %call, !dbg !38
|
|
|
|
|
br i1 %exitcond, label %if.end6, label %for.body, !dbg !38
|
|
|
|
|
; CHECK: edge for.body -> if.end6 probability is 1 / 2244 = 0.0445633%
|
|
|
|
|
; CHECK: edge for.body -> for.body probability is 2243 / 2244 = 99.9554% [HOT edge]
|
|
|
|
|
|
|
|
|
|
if.end6: ; preds = %for.body, %if.end
|
|
|
|
|
%result.0 = phi double [ 0.000000e+00, %if.end ], [ %sub, %for.body ]
|
|
|
|
|
%call7 = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([15 x i8]* @.str, i64 0, i64 0), double %result.0), !dbg !39
|
|
|
|
|
br label %return, !dbg !40
|
|
|
|
|
; CHECK: edge if.end6 -> return probability is 16 / 16 = 100% [HOT edge]
|
|
|
|
|
|
|
|
|
|
return: ; preds = %entry, %if.end6
|
|
|
|
|
%retval.0 = phi i32 [ 0, %if.end6 ], [ 1, %entry ]
|
|
|
|
|
ret i32 %retval.0, !dbg !41
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
; Function Attrs: nounwind readonly
|
|
|
|
|
declare i32 @atoi(i8* nocapture) #1
|
|
|
|
|
|
|
|
|
|
; Function Attrs: nounwind
|
|
|
|
|
declare i32 @printf(i8* nocapture readonly, ...) #2
|
|
|
|
|
|
|
|
|
|
; Function Attrs: nounwind readnone
|
|
|
|
|
declare void @llvm.dbg.value(metadata, i64, metadata) #3
|
|
|
|
|
|
|
|
|
|
attributes #0 = { nounwind uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #1 = { nounwind readonly "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #2 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #3 = { nounwind readnone }
|
|
|
|
|
attributes #4 = { nounwind readonly }
|
|
|
|
|
|
|
|
|
|
!llvm.dbg.cu = !{!0}
|
2013-11-23 05:49:45 +08:00
|
|
|
!llvm.module.flags = !{!25, !42}
|
SampleProfileLoader pass. Initial setup.
This adds a new scalar pass that reads a file with samples generated
by 'perf' during runtime. The samples read from the profile are
incorporated and emmited as IR metadata reflecting that profile.
The profile file is assumed to have been generated by an external
profile source. The profile information is converted into IR metadata,
which is later used by the analysis routines to estimate block
frequencies, edge weights and other related data.
External profile information files have no fixed format, each profiler
is free to define its own. This includes both the on-disk representation
of the profile and the kind of profile information stored in the file.
A common kind of profile is based on sampling (e.g., perf), which
essentially counts how many times each line of the program has been
executed during the run.
The SampleProfileLoader pass is organized as a scalar transformation.
On startup, it reads the file given in -sample-profile-file to
determine what kind of profile it contains. This file is assumed to
contain profile information for the whole application. The profile
data in the file is read and incorporated into the internal state of
the corresponding profiler.
To facilitate testing, I've organized the profilers to support two file
formats: text and native. The native format is whatever on-disk
representation the profiler wants to support, I think this will mostly
be bitcode files, but it could be anything the profiler wants to
support. To do this, every profiler must implement the
SampleProfile::loadNative() function.
The text format is mostly meant for debugging. Records are separated by
newlines, but each profiler is free to interpret records as it sees fit.
Profilers must implement the SampleProfile::loadText() function.
Finally, the pass will call SampleProfile::emitAnnotations() for each
function in the current translation unit. This function needs to
translate the loaded profile into IR metadata, which the analyzer will
later be able to use.
This patch implements the first steps towards the above design. I've
implemented a sample-based flat profiler. The format of the profile is
fairly simplistic. Each sampled function contains a list of relative
line locations (from the start of the function) together with a count
representing how many samples were collected at that line during
execution. I generate this profile using perf and a separate converter
tool.
Currently, I have only implemented a text format for these profiles. I
am interested in initial feedback to the whole approach before I send
the other parts of the implementation for review.
This patch implements:
- The SampleProfileLoader pass.
- The base ExternalProfile class with the core interface.
- A SampleProfile sub-class using the above interface. The profiler
generates branch weight metadata on every branch instructions that
matches the profiles.
- A text loader class to assist the implementation of
SampleProfile::loadText().
- Basic unit tests for the pass.
Additionally, the patch uses profile information to compute branch
weights based on instruction samples.
This patch converts instruction samples into branch weights. It
does a fairly simplistic conversion:
Given a multi-way branch instruction, it calculates the weight of
each branch based on the maximum sample count gathered from each
target basic block.
Note that this assignment of branch weights is somewhat lossy and can be
misleading. If a basic block has more than one incoming branch, all the
incoming branches will get the same weight. In reality, it may be that
only one of them is the most heavily taken branch.
I will adjust this assignment in subsequent patches.
llvm-svn: 194566
2013-11-13 20:22:21 +08:00
|
|
|
!llvm.ident = !{!26}
|
|
|
|
|
|
|
|
|
|
!0 = metadata !{i32 786449, metadata !1, i32 4, metadata !"clang version 3.4 (trunk 192896) (llvm/trunk 192895)", i1 true, metadata !"", i32 0, metadata !2, metadata !2, metadata !3, metadata !2, metadata !2, metadata !""} ; [ DW_TAG_compile_unit ] [./branch.cc] [DW_LANG_C_plus_plus]
|
|
|
|
|
!1 = metadata !{metadata !"branch.cc", metadata !"."}
|
|
|
|
|
!2 = metadata !{i32 0}
|
|
|
|
|
!3 = metadata !{metadata !4}
|
|
|
|
|
!4 = metadata !{i32 786478, metadata !1, metadata !5, metadata !"main", metadata !"main", metadata !"", i32 4, metadata !6, i1 false, i1 true, i32 0, i32 0, null, i32 256, i1 true, i32 (i32, i8**)* @main, null, null, metadata !12, i32 4} ; [ DW_TAG_subprogram ] [line 4] [def] [main]
|
|
|
|
|
!5 = metadata !{i32 786473, metadata !1} ; [ DW_TAG_file_type ] [./branch.cc]
|
|
|
|
|
!6 = metadata !{i32 786453, i32 0, null, metadata !"", i32 0, i64 0, i64 0, i64 0, i32 0, null, metadata !7, i32 0, null, null, null} ; [ DW_TAG_subroutine_type ] [line 0, size 0, align 0, offset 0] [from ]
|
|
|
|
|
!7 = metadata !{metadata !8, metadata !8, metadata !9}
|
|
|
|
|
!8 = metadata !{i32 786468, null, null, metadata !"int", i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] [int] [line 0, size 32, align 32, offset 0, enc DW_ATE_signed]
|
|
|
|
|
!9 = metadata !{i32 786447, null, null, metadata !"", i32 0, i64 64, i64 64, i64 0, i32 0, metadata !10} ; [ DW_TAG_pointer_type ] [line 0, size 64, align 64, offset 0] [from ]
|
|
|
|
|
!10 = metadata !{i32 786447, null, null, metadata !"", i32 0, i64 64, i64 64, i64 0, i32 0, metadata !11} ; [ DW_TAG_pointer_type ] [line 0, size 64, align 64, offset 0] [from char]
|
|
|
|
|
!11 = metadata !{i32 786468, null, null, metadata !"char", i32 0, i64 8, i64 8, i64 0, i32 0, i32 6} ; [ DW_TAG_base_type ] [char] [line 0, size 8, align 8, offset 0, enc DW_ATE_signed_char]
|
|
|
|
|
!12 = metadata !{metadata !13, metadata !14, metadata !15, metadata !17, metadata !18, metadata !21, metadata !23}
|
|
|
|
|
!13 = metadata !{i32 786689, metadata !4, metadata !"argc", metadata !5, i32 16777220, metadata !8, i32 0, i32 0} ; [ DW_TAG_arg_variable ] [argc] [line 4]
|
|
|
|
|
!14 = metadata !{i32 786689, metadata !4, metadata !"argv", metadata !5, i32 33554436, metadata !9, i32 0, i32 0} ; [ DW_TAG_arg_variable ] [argv] [line 4]
|
|
|
|
|
!15 = metadata !{i32 786688, metadata !4, metadata !"result", metadata !5, i32 7, metadata !16, i32 0, i32 0} ; [ DW_TAG_auto_variable ] [result] [line 7]
|
|
|
|
|
!16 = metadata !{i32 786468, null, null, metadata !"double", i32 0, i64 64, i64 64, i64 0, i32 0, i32 4} ; [ DW_TAG_base_type ] [double] [line 0, size 64, align 64, offset 0, enc DW_ATE_float]
|
|
|
|
|
!17 = metadata !{i32 786688, metadata !4, metadata !"limit", metadata !5, i32 8, metadata !8, i32 0, i32 0} ; [ DW_TAG_auto_variable ] [limit] [line 8]
|
|
|
|
|
!18 = metadata !{i32 786688, metadata !19, metadata !"s", metadata !5, i32 10, metadata !16, i32 0, i32 0} ; [ DW_TAG_auto_variable ] [s] [line 10]
|
|
|
|
|
!19 = metadata !{i32 786443, metadata !1, metadata !20, i32 9, i32 0, i32 2} ; [ DW_TAG_lexical_block ] [./branch.cc]
|
|
|
|
|
!20 = metadata !{i32 786443, metadata !1, metadata !4, i32 9, i32 0, i32 1} ; [ DW_TAG_lexical_block ] [./branch.cc]
|
|
|
|
|
!21 = metadata !{i32 786688, metadata !22, metadata !"u", metadata !5, i32 11, metadata !8, i32 0, i32 0} ; [ DW_TAG_auto_variable ] [u] [line 11]
|
|
|
|
|
!22 = metadata !{i32 786443, metadata !1, metadata !19, i32 11, i32 0, i32 3} ; [ DW_TAG_lexical_block ] [./branch.cc]
|
|
|
|
|
!23 = metadata !{i32 786688, metadata !24, metadata !"x", metadata !5, i32 12, metadata !16, i32 0, i32 0} ; [ DW_TAG_auto_variable ] [x] [line 12]
|
|
|
|
|
!24 = metadata !{i32 786443, metadata !1, metadata !22, i32 11, i32 0, i32 4} ; [ DW_TAG_lexical_block ] [./branch.cc]
|
|
|
|
|
!25 = metadata !{i32 2, metadata !"Dwarf Version", i32 4}
|
|
|
|
|
!26 = metadata !{metadata !"clang version 3.4 (trunk 192896) (llvm/trunk 192895)"}
|
|
|
|
|
!27 = metadata !{i32 4, i32 0, metadata !4, null}
|
|
|
|
|
!28 = metadata !{i32 5, i32 0, metadata !29, null}
|
|
|
|
|
!29 = metadata !{i32 786443, metadata !1, metadata !4, i32 5, i32 0, i32 0} ; [ DW_TAG_lexical_block ] [./branch.cc]
|
|
|
|
|
!30 = metadata !{i32 8, i32 0, metadata !4, null} ; [ DW_TAG_imported_declaration ]
|
|
|
|
|
!31 = metadata !{metadata !32, metadata !32, i64 0}
|
|
|
|
|
!32 = metadata !{metadata !"any pointer", metadata !33, i64 0}
|
|
|
|
|
!33 = metadata !{metadata !"omnipotent char", metadata !34, i64 0}
|
|
|
|
|
!34 = metadata !{metadata !"Simple C/C++ TBAA"}
|
|
|
|
|
!35 = metadata !{i32 9, i32 0, metadata !20, null}
|
|
|
|
|
!36 = metadata !{i32 13, i32 0, metadata !24, null}
|
|
|
|
|
!37 = metadata !{i32 14, i32 0, metadata !24, null}
|
|
|
|
|
!38 = metadata !{i32 11, i32 0, metadata !22, null}
|
|
|
|
|
!39 = metadata !{i32 20, i32 0, metadata !4, null}
|
|
|
|
|
!40 = metadata !{i32 21, i32 0, metadata !4, null}
|
|
|
|
|
!41 = metadata !{i32 22, i32 0, metadata !4, null}
|
2013-11-23 05:49:45 +08:00
|
|
|
!42 = metadata !{i32 1, metadata !"Debug Info Version", i32 1}
|