diff options
author | Wenlei He <aktoon@gmail.com> | 2019-11-24 23:31:02 -0800 |
---|---|---|
committer | Wenlei He <aktoon@gmail.com> | 2019-12-05 15:57:55 -0800 |
commit | e503fd85d3ac9d3e1493a7a63bc43c6939e132cc (patch) | |
tree | b1985fb306d9a6075f3c97b58ab01407692579c4 /llvm/test | |
parent | 6e8322deb3f984467c302165149a597dbbc5e279 (diff) | |
download | bcm5719-llvm-e503fd85d3ac9d3e1493a7a63bc43c6939e132cc.tar.gz bcm5719-llvm-e503fd85d3ac9d3e1493a7a63bc43c6939e132cc.zip |
[AutoFDO] Properly merge context-sensitive profile of inlinee back to outlined function
Summary:
When sample profile loader decides not to inline a previously inlined call-site, we adjust the profile of outlined function simply by scaling up its profile counts by call-site count. This means the context-sensitive profile of that inlined instance will be thrown away. This commit try to keep context-sensitive profile for such cases:
- Instead of scaling outlined function's profile, we now properly merge the FunctionSamples of inlined instance into outlined function, including all recursively inlined profile.
- Instead of adjusting the profile for negative inline decision at the end of the sample profile loader pass, we do the profile merge right after processing each function. This change paired with top-down ordering of annotation/inline-replay (a separate diff) will make sure we recursively merge profile back before the profile is used for annotation and inline replay.
A new switch -sample-profile-merge-inlinee is added to enable the new profile merge for tuning. It should be the default behavior eventually.
Reviewers: wmi, davidxl
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70653
Diffstat (limited to 'llvm/test')
5 files changed, 114 insertions, 4 deletions
diff --git a/llvm/test/CodeGen/X86/insert-prefetch-inline.afdo b/llvm/test/CodeGen/X86/insert-prefetch-inline.afdo index 83b30f6e210..935b707ff10 100644 --- a/llvm/test/CodeGen/X86/insert-prefetch-inline.afdo +++ b/llvm/test/CodeGen/X86/insert-prefetch-inline.afdo @@ -1,4 +1,4 @@ caller:0:0 - 2:sum:0 + 2: sum:0 3: 0 __prefetch_nta_0:23456 3.1: 0 __prefetch_nta_0:8764 __prefetch_nta_1:64
\ No newline at end of file diff --git a/llvm/test/Transforms/SampleProfile/Inputs/einline.prof b/llvm/test/Transforms/SampleProfile/Inputs/einline.prof index 624990b47ef..1b0f3922658 100644 --- a/llvm/test/Transforms/SampleProfile/Inputs/einline.prof +++ b/llvm/test/Transforms/SampleProfile/Inputs/einline.prof @@ -1,7 +1,7 @@ _Z3foov:200:100 - 1: _Z3barv:0 + 1: _ZL3barv:0 2: no_inline:100 - 3: _Z3barv:100 + 3: _ZL3barv:100 recursive:200:100 1: recursive:100 2: recursive:100 diff --git a/llvm/test/Transforms/SampleProfile/Inputs/inline-callee-update.prof b/llvm/test/Transforms/SampleProfile/Inputs/inline-callee-update.prof index b33f2bf998b..ecd93127472 100644 --- a/llvm/test/Transforms/SampleProfile/Inputs/inline-callee-update.prof +++ b/llvm/test/Transforms/SampleProfile/Inputs/inline-callee-update.prof @@ -4,7 +4,7 @@ sample_loader_inlinee:3000:0 1: direct_leaf_func:35000 11: 3000 test_cgscc_inline:63067:0 - 1: sample_loader_inlinee:1 + 1: cgscc_inlinee:1 cgscc_inlinee:3000:0 1: direct_leaf_func:35000 11: 3000
\ No newline at end of file diff --git a/llvm/test/Transforms/SampleProfile/Inputs/inline-mergeprof.prof b/llvm/test/Transforms/SampleProfile/Inputs/inline-mergeprof.prof new file mode 100644 index 00000000000..96ac2decb7c --- /dev/null +++ b/llvm/test/Transforms/SampleProfile/Inputs/inline-mergeprof.prof @@ -0,0 +1,13 @@ +main:225715:0 + 2.1: 5553 + 3: 5391 + 3.1: _Z3sumii:46 + 1: 23 + 2: _Z3subii:2 + 1: 2 + 3: 21 + +_Z3sumii:11:22 + 1: 11 + 2: 10 _Z3subii:10 + 3: 1
\ No newline at end of file diff --git a/llvm/test/Transforms/SampleProfile/inline-mergeprof.ll b/llvm/test/Transforms/SampleProfile/inline-mergeprof.ll new file mode 100644 index 00000000000..8b5989f7a62 --- /dev/null +++ b/llvm/test/Transforms/SampleProfile/inline-mergeprof.ll @@ -0,0 +1,97 @@ +; Test we lose details of not inlined profile without '-sample-profile-merge-inlinee' +; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline-mergeprof.prof -S | FileCheck -check-prefix=SCALE %s +; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline-mergeprof.prof -S | FileCheck -check-prefix=SCALE %s + +; Test we properly merge not inlined profile properly with '-sample-profile-merge-inlinee' +; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline-mergeprof.prof -sample-profile-merge-inlinee -S | FileCheck -check-prefix=MERGE %s +; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline-mergeprof.prof -sample-profile-merge-inlinee -S | FileCheck -check-prefix=MERGE %s + +@.str = private unnamed_addr constant [11 x i8] c"sum is %d\0A\00", align 1 + +define i32 @main() !dbg !6 { +entry: + %retval = alloca i32, align 4 + %s = alloca i32, align 4 + %i = alloca i32, align 4 + %tmp = load i32, i32* %i, align 4, !dbg !8 + %tmp1 = load i32, i32* %s, align 4, !dbg !8 + %call = call i32 @_Z3sumii(i32 %tmp, i32 %tmp1), !dbg !8 +; SCALE: call i32 @_Z3sumii +; MERGE: call i32 @_Z3sumii + store i32 %call, i32* %s, align 4, !dbg !8 + ret i32 0, !dbg !11 +} + +define i32 @_Z3sumii(i32 %x, i32 %y) !dbg !12 { +entry: + %x.addr = alloca i32, align 4 + %y.addr = alloca i32, align 4 + store i32 %x, i32* %x.addr, align 4 + store i32 %y, i32* %y.addr, align 4 + %tmp = load i32, i32* %x.addr, align 4, !dbg !13 + %tmp1 = load i32, i32* %y.addr, align 4, !dbg !13 + %add = add nsw i32 %tmp, %tmp1, !dbg !13 + %tmp2 = load i32, i32* %x.addr, align 4, !dbg !13 + %tmp3 = load i32, i32* %y.addr, align 4, !dbg !13 + %cmp1 = icmp ne i32 %tmp3, 100, !dbg !13 + br i1 %cmp1, label %if.then, label %if.else, !dbg !13 + +if.then: ; preds = %entry + %call = call i32 @_Z3subii(i32 %tmp2, i32 %tmp3), !dbg !14 + ret i32 %add, !dbg !14 + +if.else: ; preds = %entry + ret i32 %add, !dbg !15 +} + +define i32 @_Z3subii(i32 %x, i32 %y) !dbg !16 { +entry: + %x.addr = alloca i32, align 4 + %y.addr = alloca i32, align 4 + store i32 %x, i32* %x.addr, align 4 + store i32 %y, i32* %y.addr, align 4 + %tmp = load i32, i32* %x.addr, align 4, !dbg !17 + %tmp1 = load i32, i32* %y.addr, align 4, !dbg !17 + %add = sub nsw i32 %tmp, %tmp1, !dbg !17 + ret i32 %add, !dbg !18 +} + +declare i32 @printf(i8*, ...) + +!llvm.dbg.cu = !{!0} +!llvm.module.flags = !{!3, !4} +!llvm.ident = !{!5} + +!0 = distinct !DICompileUnit(language: DW_LANG_C_plus_plus, file: !1, producer: "clang version 3.5 ", isOptimized: false, runtimeVersion: 0, emissionKind: NoDebug, enums: !2, retainedTypes: !2, globals: !2, imports: !2) +!1 = !DIFile(filename: "calls.cc", directory: ".") +!2 = !{} +!3 = !{i32 2, !"Dwarf Version", i32 4} +!4 = !{i32 1, !"Debug Info Version", i32 3} +!5 = !{!"clang version 3.5 "} +!6 = distinct !DISubprogram(name: "main", scope: !1, file: !1, line: 7, type: !7, scopeLine: 7, virtualIndex: 6, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition, unit: !0, retainedNodes: !2) +!7 = !DISubroutineType(types: !2) +!8 = !DILocation(line: 10, scope: !9) +!9 = !DILexicalBlockFile(scope: !10, file: !1, discriminator: 2) +!10 = distinct !DILexicalBlock(scope: !6, file: !1, line: 10) +!11 = !DILocation(line: 12, scope: !6) +!12 = distinct !DISubprogram(name: "sum", scope: !1, file: !1, line: 3, type: !7, scopeLine: 3, virtualIndex: 6, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition, unit: !0, retainedNodes: !2) +!13 = !DILocation(line: 4, scope: !12) +!14 = !DILocation(line: 5, scope: !12) +!15 = !DILocation(line: 6, scope: !12) +!16 = distinct !DISubprogram(name: "sub", scope: !1, file: !1, line: 20, type: !7, scopeLine: 20, virtualIndex: 6, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition, unit: !0, retainedNodes: !2) +!17 = !DILocation(line: 20, scope: !16) +!18 = !DILocation(line: 21, scope: !16) + +; SCALE: name: "sum" +; SCALE-NEXT: {!"function_entry_count", i64 46} +; SCALE: !{!"branch_weights", i32 11, i32 2} +; SCALE: !{!"branch_weights", i64 20} +; SCALE: name: "sub" +; SCALE-NEXT: {!"function_entry_count", i64 -1} + +; MERGE: name: "sum" +; MERGE-NEXT: {!"function_entry_count", i64 46} +; MERGE: !{!"branch_weights", i32 11, i32 23} +; MERGE: !{!"branch_weights", i32 10} +; MERGE: name: "sub" +; MERGE-NEXT: {!"function_entry_count", i64 3}
\ No newline at end of file |