cmd/compile: pgo causes worse performance on some benchmarks

[qiulaidongfeng]

Go version

go version devel go1.24-f428c7b7 Sat Aug 3 05:06:40 2024 +0000 windows/amd64

Output of go env in your module/workspace:

set GO111MODULE=auto
set GOARCH=amd64
set GOBIN=
set GOCACHE=D:\file\go-build
set GOENV=C:\Users\26454\AppData\Roaming\go\env
set GOEXE=.exe
set GOEXPERIMENT=
set GOFLAGS=
set GOHOSTARCH=amd64
set GOHOSTOS=windows
set GOINSECURE=
set GOMODCACHE=D:\file\gofile\pkg\mod
set GONOPROXY=
set GONOSUMDB=
set GOOS=windows
set GOPATH=D:\file\gofile
set GOPRIVATE=
set GOPROXY=https://goproxy.cn,direct
set GOROOT=C:\Users\26454\.go\current
set GOSUMDB=sum.golang.org
set GOTMPDIR=
set GOTOOLCHAIN=local
set GOTOOLDIR=C:\Users\26454\.go\current\pkg\tool\windows_amd64
set GOVCS=
set GOVERSION=devel go1.24-f428c7b7 Sat Aug 3 05:06:40 2024 +0000
set GODEBUG=
set GOTELEMETRY=on
set GOTELEMETRYDIR=C:\Users\26454\AppData\Roaming\go\telemetry
set GCCGO=gccgo
set GOAMD64=v3
set AR=ar
set CC=gcc
set CXX=g++
set CGO_ENABLED=1
go: stripping unprintable or unescapable characters from %"GOMOD"%
set GOMOD=D:\file\gofile\U������\u\u-language\ucom\go.mod
set GOWORK=
set CGO_CFLAGS=-O2 -g
set CGO_CPPFLAGS=
set CGO_CXXFLAGS=-O2 -g
set CGO_FFLAGS=-O2 -g
set CGO_LDFLAGS=-O2 -g
set PKG_CONFIG=pkg-config
set GOGCCFLAGS=-m64 -mthreads -Wl,--no-gc-sections -fmessage-length=0 -ffile-prefix-map=C:\Users\26454\AppData\Local\Temp\go-build1658404967=/tmp/go-build -gno-record-gcc-switches

What did you do?

Repeat step:

git clone https://gitee.com/u-language/u-language.git -b pgo2024-8-9 cd u-language Set environment variable URoot to the current path cd ucom

go test -run=ssss -bench=Mode2 -count=10 -pgo off > new.txt && sleep 20 && go test -run=ssss -bench=Mode2 -count=10 > pgo.txt && benchstat new.txt pgo.txt

What did you see happen?

pgo does not cause performance degradation

What did you expect to see?

goos: windows
goarch: amd64
pkg: gitee.com/u-language/u-language/ucom
cpu: AMD Ryzen 7 7840HS w/ Radeon 780M Graphics
                                                           │   new.txt   │               pgo.txt                │
                                                           │   sec/op    │    sec/op     vs base                │
ComplierToCBuildMode2/输出456-16                               4.831µ ± 1%   4.695µ ± 12%        ~ (p=0.143 n=10)
ComplierToCBuildMode2/for加1，1万次，输出10000-16                   4.264µ ± 1%   4.244µ ± 11%        ~ (p=0.684 n=10)
ComplierToCBuildMode2/for代码块和main代码块都有变量i，输出10000-16         4.881µ ± 0%   4.777µ ±  1%   -2.13% (p=0.000 n=10)
ComplierToCBuildMode2/输出1-10000中偶数的数量,比较"s"=="a"输出false-16   6.770µ ± 0%   6.649µ ±  1%   -1.79% (p=0.001 n=10)
ComplierToCBuildMode2/使用指针循环加一个int变量自增10000次-16              4.346µ ± 1%   4.244µ ±  1%   -2.34% (p=0.000 n=10)
ComplierToCBuildMode2/输出结构体字段，结果字符串11-16                     4.766µ ± 1%   5.032µ ±  8%        ~ (p=0.393 n=10)
ComplierToCBuildMode2/整数和浮点数互相转换-16                          2.953µ ± 1%   2.950µ ±  2%        ~ (p=0.927 n=10)
**ComplierToCBuildMode2/调用malloc和free-16                       23.37µ ± 1%   26.95µ ±  1%  +15.29% (p=0.000 n=10)**
ComplierToCBuildMode2/使用自操作语句-16                             6.634µ ± 3%   6.391µ ±  2%   -3.66% (p=0.000 n=10)
ComplierToCBuildMode2/使用nil-16                               5.929µ ± 3%   5.758µ ±  1%   -2.88% (p=0.000 n=10)
**ComplierToCBuildMode2/使用选择器取地址与解引用-16                        25.35µ ± 1%   28.84µ ±  1%  +13.78% (p=0.000 n=10)
ComplierToCBuildMode2/使用init函数-16                            25.95µ ± 2%   29.33µ ±  1%  +13.02% (p=0.000 n=10)**
ComplierToCBuildMode2/switch-16                              7.643µ ± 1%   7.513µ ±  1%   -1.70% (p=0.000 n=10)
ComplierToCBuildMode2/位与，位或，异或，逻辑运算,括号表达式-16                 7.216µ ± 2%   7.102µ ±  1%   -1.58% (p=0.000 n=10)
ComplierToCBuildMode2/数组类型与索引表达式-16                          6.754µ ± 1%   6.451µ ±  1%   -4.49% (p=0.000 n=10)
ComplierToCBuildMode2/goto语句-16                              2.148µ ± 2%   2.074µ ±  3%   -3.45% (p=0.008 n=10)
geomean                                                      6.784µ        6.857µ         +1.07%

                                                           │   new.txt    │                 pgo.txt                 │
                                                           │     B/s      │      B/s       vs base                  │
ComplierToCBuildMode2/输出456-16                               205.1Ki ± 0%   205.1Ki ± 10%        ~ (p=0.211 n=10)
ComplierToCBuildMode2/for加1，1万次，输出10000-16                   224.6Ki ± 4%   229.5Ki ± 11%        ~ (p=0.592 n=10)
ComplierToCBuildMode2/for代码块和main代码块都有变量i，输出10000-16         195.3Ki ± 5%   205.1Ki ±  0%   +5.00% (p=0.003 n=10)
ComplierToCBuildMode2/输出1-10000中偶数的数量,比较"s"=="a"输出false-16   146.5Ki ± 0%   146.5Ki ±  0%        ~ (p=1.000 n=10) ¹
ComplierToCBuildMode2/使用指针循环加一个int变量自增10000次-16              224.6Ki ± 0%   234.4Ki ±  4%   +4.35% (p=0.003 n=10)
ComplierToCBuildMode2/输出结构体字段，结果字符串11-16                     205.1Ki ± 0%   195.3Ki ± 10%        ~ (p=0.180 n=10)
ComplierToCBuildMode2/整数和浮点数互相转换-16                          332.0Ki ± 0%   332.0Ki ±  3%        ~ (p=1.000 n=10)
ComplierToCBuildMode2/调用malloc和free-16                       39.06Ki ± 0%   39.06Ki ±  0%        ~ (p=1.000 n=10) ¹
ComplierToCBuildMode2/使用自操作语句-16                             146.5Ki ± 0%   156.2Ki ±  6%   +6.67% (p=0.001 n=10)
ComplierToCBuildMode2/使用nil-16                               166.0Ki ± 6%   166.0Ki ±  0%        ~ (p=0.263 n=10)
ComplierToCBuildMode2/使用选择器取地址与解引用-16                        39.06Ki ± 0%   29.30Ki ±  0%  -25.00% (p=0.000 n=10)
ComplierToCBuildMode2/使用init函数-16                            39.06Ki ± 0%   29.30Ki ±  0%  -25.00% (p=0.000 n=10)
ComplierToCBuildMode2/switch-16                              127.0Ki ± 0%   127.0Ki ±  0%        ~ (p=1.000 n=10)
ComplierToCBuildMode2/位与，位或，异或，逻辑运算,括号表达式-16                 136.7Ki ± 0%   136.7Ki ±  0%        ~ (p=1.000 n=10) ¹
ComplierToCBuildMode2/数组类型与索引表达式-16                          146.5Ki ± 0%   151.4Ki ±  3%   +3.33% (p=0.033 n=10)
ComplierToCBuildMode2/goto语句-16                              459.0Ki ± 2%   468.8Ki ±  2%   +2.13% (p=0.015 n=10)
geomean                                                      144.0Ki        140.5Ki         -2.43%
¹ all samples are equal

                                                           │   new.txt    │               pgo.txt               │
                                                           │     B/op     │     B/op      vs base               │
ComplierToCBuildMode2/输出456-16                               5.986Ki ± 0%   5.916Ki ± 0%  -1.18% (p=0.000 n=10)
ComplierToCBuildMode2/for加1，1万次，输出10000-16                   5.673Ki ± 0%   5.603Ki ± 0%  -1.24% (p=0.000 n=10)
ComplierToCBuildMode2/for代码块和main代码块都有变量i，输出10000-16         6.135Ki ± 0%   6.064Ki ± 0%  -1.15% (p=0.000 n=10)
ComplierToCBuildMode2/输出1-10000中偶数的数量,比较"s"=="a"输出false-16   7.929Ki ± 0%   7.858Ki ± 0%  -0.89% (p=0.000 n=10)
ComplierToCBuildMode2/使用指针循环加一个int变量自增10000次-16              5.728Ki ± 0%   5.657Ki ± 0%  -1.24% (p=0.000 n=10)
ComplierToCBuildMode2/输出结构体字段，结果字符串11-16                     5.735Ki ± 0%   5.664Ki ± 0%  -1.24% (p=0.000 n=10)
ComplierToCBuildMode2/整数和浮点数互相转换-16                          3.651Ki ± 0%   3.627Ki ± 0%  -0.67% (p=0.000 n=10)
ComplierToCBuildMode2/调用malloc和free-16                       8.631Ki ± 0%   8.561Ki ± 0%  -0.81% (p=0.000 n=10)
ComplierToCBuildMode2/使用自操作语句-16                             7.153Ki ± 0%   7.082Ki ± 0%  -1.00% (p=0.000 n=10)
ComplierToCBuildMode2/使用nil-16                               6.457Ki ± 0%   6.386Ki ± 0%  -1.10% (p=0.000 n=10)
ComplierToCBuildMode2/使用选择器取地址与解引用-16                        9.689Ki ± 0%   9.618Ki ± 0%  -0.73% (p=0.000 n=10)
ComplierToCBuildMode2/使用init函数-16                            10.63Ki ± 0%   10.56Ki ± 0%  -0.67% (p=0.000 n=10)
ComplierToCBuildMode2/switch-16                              8.242Ki ± 0%   8.171Ki ± 0%  -0.86% (p=0.000 n=10)
ComplierToCBuildMode2/位与，位或，异或，逻辑运算,括号表达式-16                 7.670Ki ± 0%   7.646Ki ± 0%  -0.31% (p=0.000 n=10)
ComplierToCBuildMode2/数组类型与索引表达式-16                          7.225Ki ± 0%   7.202Ki ± 0%  -0.32% (p=0.000 n=10)
ComplierToCBuildMode2/goto语句-16                              2.836Ki ± 0%   2.812Ki ± 0%  -0.83% (p=0.000 n=10)
geomean                                                      6.515Ki        6.457Ki       -0.89%

                                                           │  new.txt   │              pgo.txt              │
                                                           │ allocs/op  │ allocs/op   vs base               │
ComplierToCBuildMode2/输出456-16                               82.00 ± 0%   80.00 ± 0%  -2.44% (p=0.000 n=10)
ComplierToCBuildMode2/for加1，1万次，输出10000-16                   70.00 ± 0%   68.00 ± 0%  -2.86% (p=0.000 n=10)
ComplierToCBuildMode2/for代码块和main代码块都有变量i，输出10000-16         78.00 ± 0%   76.00 ± 0%  -2.56% (p=0.000 n=10)
ComplierToCBuildMode2/输出1-10000中偶数的数量,比较"s"=="a"输出false-16   111.0 ± 0%   109.0 ± 0%  -1.80% (p=0.000 n=10)
ComplierToCBuildMode2/使用指针循环加一个int变量自增10000次-16              76.00 ± 0%   74.00 ± 0%  -2.63% (p=0.000 n=10)
ComplierToCBuildMode2/输出结构体字段，结果字符串11-16                     81.00 ± 0%   79.00 ± 0%  -2.47% (p=0.000 n=10)
ComplierToCBuildMode2/整数和浮点数互相转换-16                          50.00 ± 0%   49.00 ± 0%  -2.00% (p=0.000 n=10)
ComplierToCBuildMode2/调用malloc和free-16                       94.00 ± 0%   92.00 ± 0%  -2.13% (p=0.000 n=10)
ComplierToCBuildMode2/使用自操作语句-16                             110.0 ± 0%   108.0 ± 0%  -1.82% (p=0.000 n=10)
ComplierToCBuildMode2/使用nil-16                               108.0 ± 0%   106.0 ± 0%  -1.85% (p=0.000 n=10)
ComplierToCBuildMode2/使用选择器取地址与解引用-16                        127.0 ± 0%   125.0 ± 0%  -1.57% (p=0.000 n=10)
ComplierToCBuildMode2/使用init函数-16                            132.0 ± 0%   130.0 ± 0%  -1.52% (p=0.000 n=10)
ComplierToCBuildMode2/switch-16                              130.0 ± 0%   128.0 ± 0%  -1.54% (p=0.000 n=10)
ComplierToCBuildMode2/位与，位或，异或，逻辑运算,括号表达式-16                 118.0 ± 0%   117.0 ± 0%  -0.85% (p=0.000 n=10)
ComplierToCBuildMode2/数组类型与索引表达式-16                          128.0 ± 0%   127.0 ± 0%  -0.78% (p=0.000 n=10)
ComplierToCBuildMode2/goto语句-16                              39.00 ± 0%   38.00 ± 0%  -2.56% (p=0.000 n=10)
geomean                                                      90.95        89.17       -1.96%

[gabyhelp]

Related Issues and Documentation

• cmd/compile: Building a tool chain using PGO does not achieve the 2-7% improvement claimed by blogs #63407 (closed)
• testing: Inconsistent benchmark data when  GOMAXPROCS=1 #31599
• cmd/dist: `GOMAXPROCS=2 runtime -cpu=1,2,4 -quick` makes poor use of CPU resources, increasing `all.bat` latency #65164 (closed)
• x/perf/benchstat: no output when parsing `go test` results on windows #61247 (closed)
• cmd/compile: poor register allocation with PGO #58298
• Benchmark results vastly different when Tests are skipped with -run=XXX #42696 (closed)
• runtime: huge performance drops in a timer application #35395 (closed)
• testing: benchmark cleanup causes inconsistent and slower benchmark time #38314 (closed)
• cmd/go: go run pkg is significantly slower than running built binary #25416 (closed)
• runtime: with -test.blockprofile program run even faster. #57319 (closed)

_{(Emoji vote if this was helpful or unhelpful; more detailed feedback welcome in this discussion.)}

[randall77]

Are these benchmarks of the compiler itself? (I'm not sure, but the names look suggestive.) If so, it is expected that pgo compiles will take a little bit longer.

[qiulaidongfeng]

Are these benchmarks of the compiler itself? (I'm not sure, but the names look suggestive.) If so, it is expected that pgo compiles will take a little bit longer.

These benckmark come from open source code that I wrote while researching compilation principles, it doesn't call go build.

I started using pgo in this open source code more than a year ago, default.pgo comes from running all benchmarks ,because it conforms to the https://go.dev/doc/pgo it is important that profiles are representative of actual behavior in the application’s production environment.

pgo usually brings some performance improvement. But yesterday, I tried to update default.pgo to test how much optimization pgo produces for new code, and I noticed that some of the benchmarks showed a stable decline in performance after using pgo, with the benchmark showing a stable increase in execution time of more than 10%. I see https://go.dev/doc/pgo If you encounter a program where PGO results in worse performance than disabling PGO, please file an issue at https://go.dev/issue/new.

[qiulaidongfeng]

I tried go1.21.7, and pgo gave me a performance boost of less than 5%. I tried go1.22.2, and pgo caused some benchmarks to increase execution time by more than 10%.

git bisect to b56645a87b28840a180d64077877cb46570b4176

[randall77]

@prattmic

[prattmic]

Does https://github.com/golang/go/commit/b56645a87b28840a180d64077877cb46570b4176 on its own cause regressions for these benchmarks (without PGO)?

We could set up bisection of the specific optimizations, but we'd need to modify the compiler (I did this once before). Before that, it may help to look at the logs of go build -gcflags=all=-d=pgodebug=1. Messages with "hot-budget" indicate which functions are PGO-inlined. You may find some important ones newly missing after https://github.com/golang/go/commit/b56645a87b28840a180d64077877cb46570b4176.

[qiulaidongfeng]

Before that, it may help to look at the logs of go build -gcflags=all=-d=pgodebug=1. Messages with "hot-budget" indicate which functions are PGO-inlined. You may find some important ones newly missing after https://github.com/golang/go/commit/b56645a87b28840a180d64077877cb46570b4176.

I tried , but the two outputs differ by tens of kilobytes，it is difficult to see directly what has changed. b56645a87b28840a180d64077877cb46570b4176.txt

657c885fb9278f03d5b18bfb7eeca98c25ef67ac.txt

[prattmic]

Thanks. If you sort the files first, then diff works pretty well, though there are quite a few changes.

Almost all of the changes are reductions in inline cost, which results in more inlines. This is expected behavior of https://github.com/golang/go/commit/b56645a87b28840a180d64077877cb46570b4176, so it doesn't look like something is fundamentally broken.

Bisecting the optimization decisions is a good next step. I'll put together a CL for that when I get a chance.

In the meantime, I'm curious how "diverse" the regressed benchmarks are. For example, if you look at a CPU profile of just ComplierToCBuildMode2/使用选择器取地址与解引用 is most CPU time spent spent in a single function, or is it spread across lots of functions in the program? (We can get a sense of this by taking a high-level look at a profile's flame graph).

[qiulaidongfeng]

Does https://github.com/golang/go/commit/b56645a87b28840a180d64077877cb46570b4176 on its own cause regressions for these benchmarks (without PGO)?

I tried, result discover no performance degradation without pgo.

                                                           │   old.txt   │              new.txt               │
                                                           │   sec/op    │   sec/op     vs base               │
ComplierToCBuildMode2/输出456-16                               4.788µ ± 1%   4.755µ ± 1%       ~ (p=0.078 n=10)
ComplierToCBuildMode2/for加1，1万次，输出10000-16                   4.225µ ± 2%   4.196µ ± 1%       ~ (p=0.138 n=10)
ComplierToCBuildMode2/for代码块和main代码块都有变量i，输出10000-16         4.830µ ± 1%   4.838µ ± 1%       ~ (p=0.754 n=10)
ComplierToCBuildMode2/输出1-10000中偶数的数量,比较"s"=="a"输出false-16   6.735µ ± 2%   6.719µ ± 2%       ~ (p=0.315 n=10)
ComplierToCBuildMode2/使用指针循环加一个int变量自增10000次-16              4.294µ ± 2%   4.283µ ± 2%       ~ (p=0.956 n=10)
ComplierToCBuildMode2/输出结构体字段，结果字符串11-16                     4.774µ ± 2%   4.749µ ± 1%       ~ (p=0.093 n=10)
ComplierToCBuildMode2/整数和浮点数互相转换-16                          2.947µ ± 1%   2.940µ ± 1%       ~ (p=0.699 n=10)
ComplierToCBuildMode2/调用malloc和free-16                       22.83µ ± 0%   22.85µ ± 1%       ~ (p=0.631 n=10)
ComplierToCBuildMode2/使用自操作语句-16                             6.516µ ± 2%   6.599µ ± 2%       ~ (p=0.218 n=10)
ComplierToCBuildMode2/使用nil-16                               5.873µ ± 4%   5.826µ ± 1%       ~ (p=0.159 n=10)
ComplierToCBuildMode2/使用选择器取地址与解引用-16                        24.82µ ± 1%   24.58µ ± 1%  -0.96% (p=0.009 n=10)
ComplierToCBuildMode2/使用init函数-16                            25.19µ ± 1%   25.15µ ± 0%       ~ (p=0.315 n=10)
ComplierToCBuildMode2/switch-16                              7.834µ ± 2%   7.820µ ± 1%       ~ (p=0.481 n=10)
ComplierToCBuildMode2/位与，位或，异或，逻辑运算,括号表达式-16                 7.167µ ± 1%   7.212µ ± 1%       ~ (p=0.393 n=10)
ComplierToCBuildMode2/数组类型与索引表达式-16                          6.737µ ± 2%   6.683µ ± 1%       ~ (p=0.247 n=10)
ComplierToCBuildMode2/goto语句-16                              2.135µ ± 2%   2.122µ ± 1%       ~ (p=0.382 n=10)
geomean                                                      6.725µ        6.709µ       -0.25%

[qiulaidongfeng]

I wrote an open source tool(https://github.com/qiulaidongfeng/diffbench) to analyze pgo debug information, I got:

// old inline , but new not inline
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*VarNode).C (cost 1217) at gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13 in function gitee.com/u-language/u-language/ucom/cast2.generateCFile
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.ParserVAR (cost 727) at gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20 in function gitee.com/u-language/u-language/ucom/ast.ParserNode
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.parserSece (cost 1652) at gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29 in function gitee.com/u-language/u-language/ucom/ast.ParserNode
hot-budget check allows inlining for call reflect.convertOp (cost 1568) at reflect/value.go:3322:17 in function reflect.Value.Convert
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.(*Sbt).addFunc (cost 872) at gitee.com/u-language/u-language/ucom/ast/sbt.go:199:18 in function gitee.com/u-language/u-language/ucom/ast.(*Sbt).AddFunc
hot-budget check allows inlining for call runtime.(*mcache).refill (cost 1746) at runtime/malloc.go:923:11 in function runtime.(*mcache).nextFree
----------------
// new inline , but old not inline
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserType2 (cost 540) at gitee.com/u-language/u-language/ucom/ast2/parser.go:220:22 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserType
hot-budget check allows inlining for call runtime.(*mcentral).cacheSpan (cost 1698) at runtime/mcache.go:182:44 in function runtime.(*mcache).refill
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/check.check_selector (cost 1871) at gitee.com/u-language/u-language/ucom/check/ret.go:88:57 in function gitee.com/u-language/u-language/ucom/check.ret_info
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).clonePtr (cost 1426) at github.com/huandu/go-clone@v1.7.2/clone.go:106:24 in function github.com/huandu/go-clone.(*cloneState).clone     
hot-budget check allows inlining for call runtime.gcDrainMarkWorkerIdle (cost 962) at runtime/mgc.go:1395:26 in function runtime.gcBgMarkWorker.func2
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/cast.generateInit (cost 1710) at gitee.com/u-language/u-language/ucom/cast/ast.go:144:22 in function gitee.com/u-language/u-language/ucom/cast.(*UtoC).C
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserGoto (cost 216) at gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parser
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/check3.checkCall (cost 1113) at gitee.com/u-language/u-language/ucom/check3/check3.go:45:39 in function gitee.com/u-language/u-language/ucom/check3.checkNode
hot-budget check allows inlining for call reflect.implements (cost 1846) at reflect/value.go:3552:15 in function reflect.convertOp
hot-budget check allows inlining for call runtime.markrootSpans (cost 652) at runtime/mgcmark.go:193:16 in function runtime.markroot
hot-budget check allows inlining for call runtime.scanframeworker (cost 1587) at runtime/mgcmark.go:803:18 in function runtime.scanstack
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).cloneSlice (cost 1419) at github.com/huandu/go-clone@v1.7.2/clone.go:108:26 in function github.com/huandu/go-clone.(*cloneState).clone   
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.(*Sbt).AddMethod (cost 468) at gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25 in function gitee.com/u-language/u-language/ucom/ast.parserMethod
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserObject3 (cost 324) at gitee.com/u-language/u-language/ucom/ast2/parser.go:501:25 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserSelect2
hot-budget check allows inlining for call reflect.haveIdenticalUnderlyingType (cost 1974) at reflect/value.go:3541:32 in function reflect.convertOp
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).cloneInterface (cost 1607) at github.com/huandu/go-clone@v1.7.2/clone.go:102:30 in function github.com/huandu/go-clone.(*cloneState).clone
hot-budget check allows inlining for call runtime.procresize (cost 1620) at runtime/proc.go:1452:18 in function runtime.startTheWorldWithSema
hot-budget check allows inlining for call runtime.(*pageAlloc).allocRange (cost 1012) at runtime/mpagealloc.go:914:21 in function runtime.(*pageAlloc).alloc
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.parserASSIGN (cost 1865) at gitee.com/u-language/u-language/ucom/ast/Parser.go:250:25 in function gitee.com/u-language/u-language/ucom/ast.ParserVAR
hot-budget check allows inlining for call runtime.finishsweep_m (cost 1312) at runtime/mgc.go:684:16 in function runtime.gcStart.func2
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).findElse (cost 533) at gitee.com/u-language/u-language/ucom/ast2/parser.go:557:19 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserIf
hot-budget check allows inlining for call runtime.gcDrain (cost 958) at runtime/mgcmark.go:1017:9 in function runtime.gcDrainMarkWorkerIdle
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/check3.predefined_func (cost 568) at gitee.com/u-language/u-language/ucom/check3/check3.go:214:31 in function gitee.com/u-language/u-language/ucom/check3.checkCall
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.parserASSIGN (cost 1865) at gitee.com/u-language/u-language/ucom/ast/Parser.go:478:22 in function gitee.com/u-language/u-language/ucom/ast.parserSece
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).copyStruct (cost 1558) at github.com/huandu/go-clone@v1.7.2/clone.go:230:19 in function github.com/huandu/go-clone.(*cloneState).clonePtr
hot-budget check allows inlining for call runtime.gcDrain (cost 958) at runtime/mgcmark.go:1027:9 in function runtime.gcDrainMarkWorkerDedicated
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/cast2.generateInit (cost 1817) at gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22 in function gitee.com/u-language/u-language/ucom/cast2.(*UtoC).C
hot-budget check allows inlining for call runtime.suspendG (cost 1157) at runtime/mgcmark.go:232:23 in function runtime.markroot.func1
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserIndexExprOrGenericInstantiation (cost 890) at gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55 in function gitee.com/u-language/u-language/ucom/ast2.(*stack).AutoPush
hot-budget check allows inlining for call runtime.gcDrainN (cost 1293) at runtime/mgcmark.go:568:22 in function runtime.gcAssistAlloc1
hot-budget check allows inlining for call runtime.newproc1 (cost 1681) at runtime/proc.go:4492:19 in function runtime.newproc.func1
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/cast.(*Package).GenerateheaderFile (cost 947) at gitee.com/u-language/u-language/ucom/cast/package.go:188:36 in function gitee.com/u-language/u-language/ucom/cast.(*Package).C
hot-budget check allows inlining for call runtime.(*pageAlloc).free (cost 1767) at runtime/mheap.go:1640:14 in function runtime.(*mheap).freeSpanLocked

[qiulaidongfeng]

I studied the code paths that were affected,in hot path https://gitee.com/u-language/u-language/blob/pgo2024-8-9/ucom/check3/check3.go#L231, there executed through multiple function calls, execute to errcode.NewMsgUnknownSymbol in https://gitee.com/u-language/u-language/blob/pgo2024-8-9/ucom/check3/check3.go#L127, remove https://gitee.com/u-language/u-language/blob/pgo2024-8-9/ucom/errcode/msg.go#L149-L151 of errcode.NewMsgUnknownSymbol ,no significant performance degradation with pgo.

I don't know why having these three lines of code would cause performance to drop when using pgo.

if Debug { msg.stack = debug.Stack() }

[prattmic]

Interesting. My best guess is that the additional PGO inlines made a caller of NewMsgUnknownSymbol too expensive to inline into. The "hot-budget" logs only inlines explicitly allowed by PGO. There could be some downstream effects on non-PGO inlines [1].

-m will show standard inlines as well. Use as go build "-gcflags=all=-d=pgodebug=1 -m". Perhaps you could update that nice diff tool to compare the "inlining call to ..." lines.

As for why inling NewMsgUnknownSymbol would matter, my best guess would be that some hot callsite manages to keep the name string on the stack if the call is inlined.

Also, for what it's worth, I don't see any obvious code changing errcode.Debug. If you can make it a const, this problem will likely go away as that conditional code will disappear.

[1] For example, we won't inline into a function that is very large. And inlines into a function may make it large.

[qiulaidongfeng]

-m will show standard inlines as well. Use as go build "-gcflags=all=-d=pgodebug=1 -m". Perhaps you could update that nice diff tool to compare the "inlining call to ..." lines.

I updated diff tool support inline and escape analysis information. I got:

// old inline , but new not inline
:1: inlining call to srcFunc.name
gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20: ... argument does not escape
gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20: inlining call to (*Tree).Panic
gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20: inlining call to ParserVAR
gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20: inlining call to check_is_type_token
gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20: inlining call to lex.NewLine
gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29: &SelfOpStmt{...} escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29: ... argument does not escape
gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29: inlining call to (*Tree).Panic
gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29: inlining call to NewSelfOpStmt
gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29: inlining call to parserSece
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1145:25: UTF16ToString(ret[:n]) escapes to heap
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.(*Sbt).addFunc (cost 872) at gitee.com/u-language/u-language/ucom/ast/sbt.go:199:18 in function gitee.com/u-language/u-language/ucom/ast.(*Sbt).AddFunc
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.ParserVAR (cost 727) at gitee.com/u-language/u-language/ucom/ast/Parser.go:66:20 in function gitee.com/u-language/u-language/ucom/ast.ParserNode
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.parserSece (cost 1652) at gitee.com/u-language/u-language/ucom/ast/Parser.go:78:29 in function gitee.com/u-language/u-language/ucom/ast.ParserNode
hot-budget check allows inlining for call reflect.convertOp (cost 1568) at reflect/value.go:3322:17 in function reflect.Value.Convert
hot-budget check allows inlining for call runtime.(*mcache).refill (cost 1746) at runtime/malloc.go:923:11 in function runtime.(*mcache).nextFree
reflect/value.go:3322:17: inlining call to abi.(*Type).Elem
reflect/value.go:3322:17: inlining call to abi.(*Type).Kind
reflect/value.go:3322:17: inlining call to convertOp
reflect/value.go:3322:17: inlining call to nameFor
reflect/value.go:3322:17: inlining call to pkgPathFor
reflect/value.go:3322:17: inlining call to toRType
runtime/debugcall.go:55:19: inlining call to funcInfo.valid
runtime/debugcall.go:55:19: inlining call to funcname
runtime/debuglog.go:826:19: inlining call to funcInfo.valid
runtime/debuglog.go:826:19: inlining call to funcname
runtime/error.go:307:35: inlining call to funcInfo.valid
runtime/error.go:307:35: inlining call to funcname
runtime/heapdump.go:288:18: inlining call to funcInfo.valid
runtime/heapdump.go:288:18: inlining call to funcname
runtime/heapdump.go:634:20: inlining call to funcInfo.valid
runtime/heapdump.go:634:20: inlining call to funcname
runtime/malloc.go:923:11: inlining call to (*mcache).refill
runtime/malloc.go:923:11: inlining call to atomic.(*Uint64).Add
runtime/malloc.go:923:11: inlining call to spanClass.sizeclass
runtime/mgc.go:1376:31: inlining call to gcDrainMarkWorkerDedicated
runtime/mgc.go:1391:31: inlining call to gcDrainMarkWorkerDedicated
runtime/mgcmark.go:802:38: inlining call to funcname
runtime/os_windows.go:1329:35: inlining call to srcFunc.name
runtime/panic.go:56:45: inlining call to funcInfo.valid
runtime/panic.go:56:45: inlining call to funcname
runtime/plugin.go:105:19: inlining call to funcInfo.valid
runtime/plugin.go:105:19: inlining call to funcname
runtime/plugin.go:114:20: inlining call to funcInfo.valid
runtime/plugin.go:114:20: inlining call to funcname
runtime/preempt.go:418:28: inlining call to srcFunc.name
runtime/print.go:295:23: inlining call to funcInfo.valid
runtime/print.go:295:23: inlining call to funcname
runtime/proc.go:3824:60: inlining call to funcInfo.valid
runtime/proc.go:3824:60: inlining call to funcname
runtime/stack.go:625:53: inlining call to funcInfo.valid
runtime/stack.go:625:53: inlining call to funcname
runtime/stack.go:930:38: inlining call to funcname
runtime/stack.go:985:21: inlining call to funcInfo.valid
runtime/stack.go:985:21: inlining call to funcname
runtime/stkframe.go:101:17: inlining call to funcInfo.valid
runtime/stkframe.go:101:17: inlining call to funcname
runtime/stkframe.go:127:44: inlining call to funcInfo.valid
runtime/stkframe.go:127:44: inlining call to funcname
runtime/stkframe.go:139:43: inlining call to funcInfo.valid
runtime/stkframe.go:139:43: inlining call to funcname
runtime/stkframe.go:194:37: inlining call to funcInfo.valid
runtime/stkframe.go:194:37: inlining call to funcname
runtime/stkframe.go:201:104: inlining call to funcInfo.valid
runtime/stkframe.go:201:104: inlining call to funcname
runtime/stkframe.go:221:37: inlining call to funcInfo.valid
runtime/stkframe.go:221:37: inlining call to funcname
runtime/stkframe.go:226:102: inlining call to funcInfo.valid
runtime/stkframe.go:226:102: inlining call to funcname
runtime/symtab.go:1017:22: inlining call to funcname
runtime/symtab.go:1018:19: inlining call to funcname
runtime/symtab.go:1032:17: inlining call to funcname
runtime/symtab.go:1067:17: inlining call to funcname
runtime/symtab.go:1075:17: inlining call to funcname
runtime/symtab.go:132:39: inlining call to srcFunc.name
runtime/symtab.go:185:16: inlining call to srcFunc.name
runtime/symtab.go:544:22: inlining call to funcInfo.valid
runtime/symtab.go:544:22: inlining call to funcname
runtime/symtab.go:546:99: inlining call to funcInfo.valid
runtime/symtab.go:546:99: inlining call to funcname
runtime/symtab.go:548:56: inlining call to funcInfo.valid
runtime/symtab.go:548:56: inlining call to funcname
runtime/symtab.go:669:21: inlining call to srcFunc.name
runtime/symtab.go:687:34: inlining call to funcInfo.valid
runtime/symtab.go:687:34: inlining call to funcname
runtime/symtab.go:958:56: inlining call to funcInfo.valid
runtime/symtab.go:958:56: inlining call to funcname
runtime/symtab.go:984:35: inlining call to funcInfo.valid
runtime/symtab.go:984:35: inlining call to funcname
runtime/traceback.go:1001:19: inlining call to srcFunc.name
runtime/traceback.go:1090:34: inlining call to srcFunc.name
runtime/traceback.go:1141:17: inlining call to srcFunc.name
runtime/traceback.go:1330:27: inlining call to funcInfo.valid
runtime/traceback.go:1330:27: inlining call to funcname
runtime/traceback.go:223:19: inlining call to funcInfo.valid
runtime/traceback.go:223:19: inlining call to funcname
runtime/traceback.go:364:62: inlining call to funcInfo.valid
runtime/traceback.go:364:62: inlining call to funcname
runtime/traceback.go:468:73: inlining call to funcInfo.valid
runtime/traceback.go:468:73: inlining call to funcname
runtime/traceback.go:515:19: inlining call to funcInfo.valid
runtime/traceback.go:515:19: inlining call to funcname
runtime/traceback.go:625:43: inlining call to funcname
runtime/traceback.go:802:24: inlining call to funcInfo.valid
runtime/traceback.go:802:24: inlining call to funcname
----------------
// new inline , but old not inline
:1: inlining call to (*moduledata).funcName
:1: inlining call to (*mspan).newPinnerBits
:1: inlining call to Float64Slice.Sort
:1: inlining call to IntSlice.Sort
:1: inlining call to StringSlice.Sort
:1: x escapes to heap
encoding/binary/binary.go:455:23: inlining call to reflect.Indirect
encoding/binary/binary.go:455:23: inlining call to reflect.Value.Kind
encoding/binary/binary.go:455:23: inlining call to reflect.flag.kind
encoding/binary/binary.go:471:34: inlining call to reflect.Indirect
encoding/binary/binary.go:471:34: inlining call to reflect.Value.Kind
encoding/binary/binary.go:471:34: inlining call to reflect.flag.kind
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: Type + "method" escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: inlining call to (*Sbt).AddMethod
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: inlining call to NewMethodTableInfo
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: inlining call to NewSbt
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: make(map[string]SymbolInfo) escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: new(Sbt) escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25: new(sync.Map) escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:1331:21: inlining call to (*Sbt).AddEnum
gitee.com/u-language/u-language/ucom/ast/Parser.go:192:21: inlining call to (*Sbt).AddEnum
gitee.com/u-language/u-language/ucom/ast/Parser.go:250:25: &AssignmentNode{...} escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:250:25: inlining call to NewASSIGNNode
gitee.com/u-language/u-language/ucom/ast/Parser.go:250:25: inlining call to parserASSIGN
gitee.com/u-language/u-language/ucom/ast/Parser.go:324:21: inlining call to (*Sbt).AddConst
gitee.com/u-language/u-language/ucom/ast/Parser.go:454:23: inlining call to (*Sbt).addFunc
gitee.com/u-language/u-language/ucom/ast/Parser.go:478:22: &AssignmentNode{...} escapes to heap
gitee.com/u-language/u-language/ucom/ast/Parser.go:478:22: inlining call to NewASSIGNNode
gitee.com/u-language/u-language/ucom/ast/Parser.go:478:22: inlining call to parserASSIGN
gitee.com/u-language/u-language/ucom/ast/check.go:223:17: inlining call to utils.PutMapSet
gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55: "exit" escapes to heap
gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55: &IndexExpr{...} escapes to heap
gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55: ... argument does not escape
gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55: inlining call to (*Tree).panic
gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55: inlining call to (*Tree).parserIndexExprOrGenericInstantiation
gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55: inlining call to NewIndexExpr
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: "exit" escapes to heap
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: &GotoStmt{} escapes to heap
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: ... argument does not escape
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: inlining call to (*Tree).lex2TokenToStr
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: inlining call to (*Tree).panic
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: inlining call to (*Tree).parserGoto
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: inlining call to (*Tree).subline
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: inlining call to NewGotoStmt
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: inlining call to errcode.NewMsgUnexpected
gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15: ~r0 escapes to heap
gitee.com/u-language/u-language/ucom/ast2/parser.go:220:22: ... argument does not escape
gitee.com/u-language/u-language/ucom/ast2/parser.go:220:22: inlining call to (*Tree).parserType2
gitee.com/u-language/u-language/ucom/ast2/parser.go:501:25: ... argument does not escape
gitee.com/u-language/u-language/ucom/ast2/parser.go:501:25: inlining call to (*Tree).parserObject3
gitee.com/u-language/u-language/ucom/ast2/parser.go:557:19: inlining call to (*Tree).findElse
gitee.com/u-language/u-language/ucom/ast2/parser.go:557:19: inlining call to NewRbraceNode
gitee.com/u-language/u-language/ucom/ast2/parser.go:557:19: ~r0 escapes to heap
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: "strings: illegal use of non-zero Builder copied by value" escapes to heap
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: &ast.FuncInfo{} does not escape
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: inlining call to generateInit
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: inlining call to strings.(*Builder).String
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: inlining call to strings.(*Builder).WriteString
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: inlining call to strings.(*Builder).copyCheck
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast/ast.go:144:22: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast/ir2.go:155:7: inlining call to ast.(*MethodNode).C
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: "strings: illegal use of non-zero Builder copied by value" escapes to heap
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: func literal does not escape
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to (*Package).GenerateheaderFile
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to (*Package).WithAutoFree
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to strings.(*Builder).String
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to strings.(*Builder).WriteString
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to strings.(*Builder).copyCheck
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast/package.go:188:36: p.headerFile.Data escapes to heap
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: "strings: illegal use of non-zero Builder copied by value" escapes to heap
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: &ast2.FuncInfo{} does not escape
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: inlining call to generateInit
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: inlining call to strings.(*Builder).String
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: inlining call to strings.(*Builder).WriteString
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: inlining call to strings.(*Builder).copyCheck
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13: "strings: illegal use of non-zero Builder copied by value" escapes to heap
gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13: inlining call to ast2.(*VarNode).C
gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13: inlining call to strings.(*Builder).WriteString
gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13: inlining call to strings.(*Builder).copyCheck
gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13: inlining call to strings.noescape
gitee.com/u-language/u-language/ucom/check/check.go:134:23: inlining call to utils.PutMapSet
gitee.com/u-language/u-language/ucom/check/check.go:175:24: inlining call to utils.PutMapSet
gitee.com/u-language/u-language/ucom/check/check.go:27:23: inlining call to utils.PutMapSet
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: "&" + typ escapes to heap
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: &ast.Object{...} escapes to heap
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: ast.info escapes to heap
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to IsExportSymbol
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to ast.(*Sbt).Have
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to ast.(*SymbolInfo).PtrStructDecl
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to ast.(*SymbolInfo).PtrVarInfoSbt
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to ast.NewObject
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to ast.NewSymbolInfo_Var
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to ast.NewVarInfoSbt
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to check_selector
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to errcode.NewMsgSymbol
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: inlining call to errcode.remove_package_name
gitee.com/u-language/u-language/ucom/check/ret.go:88:57: ~r0 escapes to heap
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: "&" + typ escapes to heap
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: ... argument does not escape
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: inlining call to ast2.(*Tree).Panic
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: inlining call to check_parame_num
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: inlining call to errcode.NewMsgNumberOfParameNoMatch
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: inlining call to predefined_func
gitee.com/u-language/u-language/ucom/check3/check3.go:214:31: ~r0 escapes to heap
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: ... argument does not escape
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: errcode.NewMsgCallTypeIsNotEqual(decltyp, typ, i + 1) escapes to heap
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to ast2.(*Sbt).Have
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to ast2.(*Tree).Panic
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to checkCall
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to check_parame_num
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to errcode.NewMsgNumberOfParameNoMatch
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to errcode.NewMsgUnknownSymbol
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to errcode.remove_package_name
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to utils.IsPtr
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to utils.Typ_is_equal
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to utils.is_ptr_and_nil
gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: ~r0 escapes to heap
gitee.com/u-language/u-language/ucom/check3/check3.go:62:12: inlining call to checkElse
gitee.com/u-language/u-language/ucom/parser/package.go:38:29: inlining call to utils.NewWaitGroup
gitee.com/u-language/u-language/ucom/parser/package.go:83:10: ast.p.FindTree does not escape
gitee.com/u-language/u-language/ucom/parser/package.go:83:10: inlining call to ast.(*Package).Parser
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: &reflect.ValueError{...} escapes to heap
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: inlining call to (*cloneState).cloneInterface
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: inlining call to reflect.Value.IsNil
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: inlining call to reflect.Value.Type
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: inlining call to reflect.flag.kind
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:102:30: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: &reflect.ValueError{...} escapes to heap
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to (*Allocator).New
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to (*cloneState).clonePtr
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.Value.CanInterface
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.Value.IsNil
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.Value.Kind
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.Value.Type
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.flag.kind
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:106:24: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: &reflect.ValueError{...} escapes to heap
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to (*Allocator).MakeSlice
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to (*cloneState).cloneSlice
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.Value.Cap
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.Value.IsNil
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.Value.Len
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.Value.Type
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.flag.kind
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:108:26: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to (*cloneState).copyStruct
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.PointerTo
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.PtrTo
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.Value.CanAddr
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.Value.Type
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.noescape
github.com/huandu/go-clone@v1.7.2/clone.go:230:19: inlining call to reflect.toRType
github.com/spf13/cobra@v1.8.0/bash_completions.go:449:32: commandSorterByName(c.commands) escapes to heap
github.com/spf13/cobra@v1.8.0/bash_completions.go:449:32: inlining call to (*Command).Commands
github.com/spf13/cobra@v1.8.0/bash_completions.go:656:32: commandSorterByName(c.commands) escapes to heap
github.com/spf13/cobra@v1.8.0/bash_completions.go:656:32: inlining call to (*Command).Commands
github.com/spf13/cobra@v1.8.0/command.go:1250:40: commandSorterByName(c.commands) escapes to heap
github.com/spf13/cobra@v1.8.0/command.go:1250:40: inlining call to (*Command).Commands
github.com/spf13/cobra@v1.8.0/completions.go:286:26: commandSorterByName(c.commands) escapes to heap
github.com/spf13/cobra@v1.8.0/completions.go:286:26: inlining call to (*Command).Commands
github.com/spf13/cobra@v1.8.0/completions.go:455:45: commandSorterByName(c.commands) escapes to heap
github.com/spf13/cobra@v1.8.0/completions.go:455:45: inlining call to (*Command).Commands
github.com/spf13/pflag@v1.0.5/flag.go:213:11: inlining call to sort.StringSlice.Sort
github.com/spf13/pflag@v1.0.5/flag.go:213:11: sort.x escapes to heap
golang.org/x/sys@v0.15.0/windows/exec_windows.go:203:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:219:29: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:238:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:309:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:309:42: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:42:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:744:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:788:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:807:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/security_windows.go:824:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1096:24: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1126:23: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1130:25: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1145:25: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1145:25: ~r0 escapes to heap
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1161:32: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1260:23: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1311:27: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:1407:49: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:178:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:250:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:435:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:448:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:461:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:521:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:525:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:529:22: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:535:23: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/setupapi_windows.go:550:32: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/syscall_windows.go:1491:20: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/syscall_windows.go:1495:20: inlining call to UTF16ToString
golang.org/x/sys@v0.15.0/windows/syscall_windows.go:1680:22: inlining call to UTF16ToString
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.(*Sbt).AddMethod (cost 468) at gitee.com/u-language/u-language/ucom/ast/Parser.go:1237:25 in function gitee.com/u-language/u-language/ucom/ast.parserMethod
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.parserASSIGN (cost 1865) at gitee.com/u-language/u-language/ucom/ast/Parser.go:250:25 in function gitee.com/u-language/u-language/ucom/ast.ParserVAR
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast.parserASSIGN (cost 1865) at gitee.com/u-language/u-language/ucom/ast/Parser.go:478:22 in function gitee.com/u-language/u-language/ucom/ast.parserSece
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).findElse (cost 533) at gitee.com/u-language/u-language/ucom/ast2/parser.go:557:19 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserIf
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserGoto (cost 216) at gitee.com/u-language/u-language/ucom/ast2/parser.go:124:15 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parser
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserIndexExprOrGenericInstantiation (cost 890) at gitee.com/u-language/u-language/ucom/ast2/expr3.go:303:55 in function gitee.com/u-language/u-language/ucom/ast2.(*stack).AutoPush
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserObject3 (cost 324) at gitee.com/u-language/u-language/ucom/ast2/parser.go:501:25 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserSelect2
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserType2 (cost 540) at gitee.com/u-language/u-language/ucom/ast2/parser.go:220:22 in function gitee.com/u-language/u-language/ucom/ast2.(*Tree).parserType
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/ast2.(*VarNode).C (cost 1217) at gitee.com/u-language/u-language/ucom/cast2/utoc.go:82:13 in function gitee.com/u-language/u-language/ucom/cast2.generateCFile
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/cast.(*Package).GenerateheaderFile (cost 947) at gitee.com/u-language/u-language/ucom/cast/package.go:188:36 in function gitee.com/u-language/u-language/ucom/cast.(*Package).C
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/cast.generateInit (cost 1710) at gitee.com/u-language/u-language/ucom/cast/ast.go:144:22 in function gitee.com/u-language/u-language/ucom/cast.(*UtoC).C
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/cast2.generateInit (cost 1817) at gitee.com/u-language/u-language/ucom/cast2/utoc.go:35:22 in function gitee.com/u-language/u-language/ucom/cast2.(*UtoC).C
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/check.check_selector (cost 1871) at gitee.com/u-language/u-language/ucom/check/ret.go:88:57 in function gitee.com/u-language/u-language/ucom/check.ret_info
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/check3.checkCall (cost 1132) at gitee.com/u-language/u-language/ucom/check3/check3.go:45:39 in function gitee.com/u-language/u-language/ucom/check3.checkNode
hot-budget check allows inlining for call gitee.com/u-language/u-language/ucom/check3.predefined_func (cost 568) at gitee.com/u-language/u-language/ucom/check3/check3.go:214:31 in function gitee.com/u-language/u-language/ucom/check3.checkCall
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).cloneInterface (cost 1607) at github.com/huandu/go-clone@v1.7.2/clone.go:102:30 in function github.com/huandu/go-clone.(*cloneState).clone
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).clonePtr (cost 1426) at github.com/huandu/go-clone@v1.7.2/clone.go:106:24 in function github.com/huandu/go-clone.(*cloneState).clone
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).cloneSlice (cost 1419) at github.com/huandu/go-clone@v1.7.2/clone.go:108:26 in function github.com/huandu/go-clone.(*cloneState).clone
hot-budget check allows inlining for call github.com/huandu/go-clone.(*cloneState).copyStruct (cost 1558) at github.com/huandu/go-clone@v1.7.2/clone.go:230:19 in function github.com/huandu/go-clone.(*cloneState).clonePtr
hot-budget check allows inlining for call reflect.haveIdenticalUnderlyingType (cost 1974) at reflect/value.go:3541:32 in function reflect.convertOp
hot-budget check allows inlining for call reflect.implements (cost 1846) at reflect/value.go:3552:15 in function reflect.convertOp
hot-budget check allows inlining for call runtime.(*mcentral).cacheSpan (cost 1698) at runtime/mcache.go:182:44 in function runtime.(*mcache).refill
hot-budget check allows inlining for call runtime.(*pageAlloc).allocRange (cost 1012) at runtime/mpagealloc.go:914:21 in function runtime.(*pageAlloc).alloc
hot-budget check allows inlining for call runtime.(*pageAlloc).free (cost 1767) at runtime/mheap.go:1640:14 in function runtime.(*mheap).freeSpanLocked
hot-budget check allows inlining for call runtime.finishsweep_m (cost 1312) at runtime/mgc.go:684:16 in function runtime.gcStart.func2
hot-budget check allows inlining for call runtime.gcDrain (cost 958) at runtime/mgcmark.go:1017:9 in function runtime.gcDrainMarkWorkerIdle
hot-budget check allows inlining for call runtime.gcDrain (cost 958) at runtime/mgcmark.go:1027:9 in function runtime.gcDrainMarkWorkerDedicated
hot-budget check allows inlining for call runtime.gcDrainMarkWorkerIdle (cost 962) at runtime/mgc.go:1395:26 in function runtime.gcBgMarkWorker.func2
hot-budget check allows inlining for call runtime.gcDrainN (cost 1293) at runtime/mgcmark.go:568:22 in function runtime.gcAssistAlloc1
hot-budget check allows inlining for call runtime.markrootSpans (cost 652) at runtime/mgcmark.go:193:16 in function runtime.markroot
hot-budget check allows inlining for call runtime.newproc1 (cost 1681) at runtime/proc.go:4492:19 in function runtime.newproc.func1
hot-budget check allows inlining for call runtime.procresize (cost 1620) at runtime/proc.go:1452:18 in function runtime.startTheWorldWithSema
hot-budget check allows inlining for call runtime.scanframeworker (cost 1587) at runtime/mgcmark.go:803:18 in function runtime.scanstack
hot-budget check allows inlining for call runtime.suspendG (cost 1157) at runtime/mgcmark.go:232:23 in function runtime.markroot.func1
internal/intern/intern.go:66:19: inlining call to safeMap
internal/intern/intern.go:66:19: map[key]*Value{} escapes to heap
math/big/nat.go:1195:8: inlining call to putNat
math/big/nat.go:1197:9: inlining call to putNat
math/big/nat.go:486:9: inlining call to putNat
math/big/nat.go:511:8: inlining call to putNat
math/big/nat.go:607:9: inlining call to putNat
math/big/natdiv.go:512:8: inlining call to putNat
math/big/natdiv.go:620:8: inlining call to putNat
math/big/natdiv.go:710:8: inlining call to putNat
math/big/natdiv.go:743:10: inlining call to putNat
math/big/natdiv.go:746:8: inlining call to putNat
os/file_windows.go:130:21: inlining call to (*dirInfo).close
os/file_windows.go:154:18: inlining call to (*dirInfo).close
reflect/value.go:3541:32: inlining call to abi.(*FuncType).In
reflect/value.go:3541:32: inlining call to abi.(*FuncType).InSlice
reflect/value.go:3541:32: inlining call to abi.(*FuncType).NumIn
reflect/value.go:3541:32: inlining call to abi.(*FuncType).NumOut
reflect/value.go:3541:32: inlining call to abi.(*FuncType).Out
reflect/value.go:3541:32: inlining call to abi.(*FuncType).OutSlice
reflect/value.go:3541:32: inlining call to abi.(*StructField).Embedded
reflect/value.go:3541:32: inlining call to abi.(*Type).ChanDir
reflect/value.go:3541:32: inlining call to abi.(*Type).Elem
reflect/value.go:3541:32: inlining call to abi.(*Type).Key
reflect/value.go:3541:32: inlining call to abi.(*Type).Kind
reflect/value.go:3541:32: inlining call to abi.(*Type).Len
reflect/value.go:3541:32: inlining call to abi.Name.IsEmbedded
reflect/value.go:3541:32: inlining call to abi.addChecked
reflect/value.go:3541:32: inlining call to haveIdenticalUnderlyingType
reflect/value.go:3552:15: inlining call to (*interfaceType).nameOff
reflect/value.go:3552:15: inlining call to (*interfaceType).typeOff
reflect/value.go:3552:15: inlining call to (*rtype).nameOff
reflect/value.go:3552:15: inlining call to (*rtype).typeOff
reflect/value.go:3552:15: inlining call to abi.(*Type).Kind
reflect/value.go:3552:15: inlining call to abi.(*Type).Uncommon
reflect/value.go:3552:15: inlining call to abi.(*UncommonType).Methods
reflect/value.go:3552:15: inlining call to abi.Name.IsExported
reflect/value.go:3552:15: inlining call to abi.addChecked
reflect/value.go:3552:15: inlining call to implements
reflect/value.go:3552:15: inlining call to nameOffFor
reflect/value.go:3552:15: inlining call to toRType
reflect/value.go:3552:15: inlining call to typeOffFor
regexp/backtrack.go:315:18: inlining call to newBitState
regexp/backtrack.go:315:18: new(bitState) escapes to heap
regexp/exec.go:403:24: inlining call to newOnePassMachine
regexp/exec.go:403:24: new(onePassMachine) escapes to heap
runtime/arena.go:948:8: inlining call to sysMap
runtime/arena.go:948:8: inlining call to sysMapOS
runtime/debugcall.go:172:16: inlining call to casGToWaiting
runtime/heapdump.go:698:15: inlining call to casGToWaiting
runtime/iface.go:627:24: &TypeAssertionError{...} escapes to heap
runtime/iface.go:627:24: inlining call to assertE2I
runtime/iface.go:632:24: &TypeAssertionError{...} escapes to heap
runtime/iface.go:632:24: inlining call to assertE2I
runtime/malloc.go:1587:10: inlining call to sysMap
runtime/malloc.go:1587:10: inlining call to sysMapOS
runtime/mbarrier.go:211:22: inlining call to reflect_typedmemmove
runtime/mbitmap.go:583:17: inlining call to (*wbBuf).get1
runtime/mbitmap.go:626:16: inlining call to (*wbBuf).get1
runtime/mbitmap.go:658:18: inlining call to (*wbBuf).get1
runtime/mcache.go:182:44: inlining call to (*activeSweep).begin
runtime/mcache.go:182:44: inlining call to (*mcentral).cacheSpan
runtime/mcache.go:182:44: inlining call to (*mcentral).fullSwept
runtime/mcache.go:182:44: inlining call to (*mcentral).fullUnswept
runtime/mcache.go:182:44: inlining call to (*mcentral).partialSwept
runtime/mcache.go:182:44: inlining call to (*mcentral).partialUnswept
runtime/mcache.go:182:44: inlining call to atomic.(*Uint32).CompareAndSwap
runtime/mcache.go:182:44: inlining call to atomic.(*Uint32).Load
runtime/mcache.go:182:44: inlining call to spanClass.sizeclass
runtime/mcache.go:182:44: inlining call to traceEnabled
runtime/mfinal.go:249:34: &TypeAssertionError{...} escapes to heap
runtime/mfinal.go:249:34: inlining call to assertE2I
runtime/mfinal.go:467:24: inlining call to assertE2I2
runtime/mgc.go:1371:17: inlining call to casGToWaiting
runtime/mgc.go:684:16: inlining call to (*activeSweep).sweepers
runtime/mgc.go:684:16: inlining call to (*mcentral).fullUnswept
runtime/mgc.go:684:16: inlining call to (*mcentral).partialUnswept
runtime/mgc.go:684:16: inlining call to assertWorldStopped
runtime/mgc.go:684:16: inlining call to atomic.(*Uint32).Load
runtime/mgc.go:684:16: inlining call to finishsweep_m
runtime/mgc.go:833:16: inlining call to casGToWaiting
runtime/mgc.go:951:15: inlining call to casGToWaiting
runtime/mgcmark.go:1017:9: inlining call to (*gcWork).tryGetFast
runtime/mgcmark.go:1017:9: inlining call to atomic.(*Bool).Load
runtime/mgcmark.go:1017:9: inlining call to atomic.(*Int64).Add
runtime/mgcmark.go:1017:9: inlining call to atomic.(*Uint8).Load
runtime/mgcmark.go:1017:9: inlining call to gcDrain
runtime/mgcmark.go:1027:9: inlining call to (*gcWork).tryGetFast
runtime/mgcmark.go:1027:9: inlining call to atomic.(*Bool).Load
runtime/mgcmark.go:1027:9: inlining call to atomic.(*Int64).Add
runtime/mgcmark.go:1027:9: inlining call to atomic.(*Uint8).Load
runtime/mgcmark.go:1027:9: inlining call to gcDrain
runtime/mgcmark.go:193:16: inlining call to (*mSpanStateBox).get
runtime/mgcmark.go:193:16: inlining call to (*mspan).base
runtime/mgcmark.go:193:16: inlining call to arenaIdx.l1
runtime/mgcmark.go:193:16: inlining call to arenaIdx.l2
runtime/mgcmark.go:193:16: inlining call to atomic.(*Uint8).Load
runtime/mgcmark.go:193:16: inlining call to getLockRank
runtime/mgcmark.go:193:16: inlining call to lock
runtime/mgcmark.go:193:16: inlining call to lockWithRank
runtime/mgcmark.go:193:16: inlining call to markrootSpans
runtime/mgcmark.go:193:16: inlining call to spanClass.noscan
runtime/mgcmark.go:193:16: inlining call to unlock
runtime/mgcmark.go:193:16: inlining call to unlockWithRank
runtime/mgcmark.go:222:18: inlining call to casGToWaiting
runtime/mgcmark.go:232:23: func literal does not escape
runtime/mgcmark.go:232:23: inlining call to atomic.(*Uint32).Load
runtime/mgcmark.go:232:23: inlining call to dumpgstatus
runtime/mgcmark.go:232:23: inlining call to nanotime
runtime/mgcmark.go:232:23: inlining call to osyield
runtime/mgcmark.go:232:23: inlining call to readgstatus
runtime/mgcmark.go:232:23: inlining call to suspendG
runtime/mgcmark.go:563:15: inlining call to casGToWaiting
runtime/mgcmark.go:568:22: inlining call to (*gcCPULimiterState).limiting
runtime/mgcmark.go:568:22: inlining call to (*gcWork).tryGetFast
runtime/mgcmark.go:568:22: inlining call to atomic.(*Bool).Load
runtime/mgcmark.go:568:22: inlining call to atomic.(*Int64).Add
runtime/mgcmark.go:568:22: inlining call to atomic.(*Uint8).Load
runtime/mgcmark.go:568:22: inlining call to gcDrainN
runtime/mgcmark.go:802:12: inlining call to (*unwinder).init
runtime/mgcmark.go:803:18: inlining call to funcInfo.valid
runtime/mgcmark.go:803:18: inlining call to scanframeworker
runtime/mgcsweep.go:103:29: inlining call to (*spanSetBlockAlloc).free
runtime/mgcsweep.go:103:29: inlining call to atomic.(*Uint32).Store
runtime/mgcsweep.go:105:32: inlining call to (*spanSetBlockAlloc).free
runtime/mgcsweep.go:105:32: inlining call to atomic.(*Uint32).Store
runtime/mgcsweep.go:255:29: inlining call to (*spanSetBlockAlloc).free
runtime/mgcsweep.go:255:29: inlining call to atomic.(*Uint32).Store
runtime/mgcsweep.go:256:26: inlining call to (*spanSetBlockAlloc).free
runtime/mgcsweep.go:256:26: inlining call to atomic.(*Uint32).Store
runtime/mheap.go:1503:11: inlining call to sysMap
runtime/mheap.go:1503:11: inlining call to sysMapOS
runtime/mheap.go:1534:8: inlining call to sysMap
runtime/mheap.go:1534:8: inlining call to sysMapOS
runtime/mheap.go:1640:14: inlining call to (*pageAlloc).chunkOf
runtime/mheap.go:1640:14: inlining call to (*pageAlloc).free
runtime/mheap.go:1640:14: inlining call to (*pageBits).clear
runtime/mheap.go:1640:14: inlining call to (*pageBits).clearAll
runtime/mheap.go:1640:14: inlining call to (*pallocBits).free
runtime/mheap.go:1640:14: inlining call to (*pallocBits).free1
runtime/mheap.go:1640:14: inlining call to (*pallocBits).freeAll
runtime/mheap.go:1640:14: inlining call to assertLockHeld
runtime/mheap.go:1640:14: inlining call to chunkIdx.l1
runtime/mheap.go:1640:14: inlining call to chunkIdx.l2
runtime/mheap.go:1640:14: inlining call to chunkIndex
runtime/mheap.go:1640:14: inlining call to chunkPageIndex
runtime/mheap.go:1640:14: inlining call to offAddr.lessThan
runtime/mpagealloc.go:914:21: inlining call to (*pageAlloc).allocRange
runtime/mpagealloc.go:914:21: inlining call to (*pageAlloc).chunkOf
runtime/mpagealloc.go:914:21: inlining call to (*pageBits).clearAll
runtime/mpagealloc.go:914:21: inlining call to (*pageBits).setAll
runtime/mpagealloc.go:914:21: inlining call to (*pallocBits).allocAll
runtime/mpagealloc.go:914:21: inlining call to (*pallocData).allocAll
runtime/mpagealloc.go:914:21: inlining call to assertLockHeld
runtime/mpagealloc.go:914:21: inlining call to chunkIdx.l1
runtime/mpagealloc.go:914:21: inlining call to chunkIdx.l2
runtime/mpagealloc.go:914:21: inlining call to chunkIndex
runtime/mpagealloc.go:914:21: inlining call to chunkPageIndex
runtime/mpagealloc_64bit.go:179:9: inlining call to sysMap
runtime/mpagealloc_64bit.go:179:9: inlining call to sysMapOS
runtime/mpagealloc_64bit.go:235:9: inlining call to sysMap
runtime/mpagealloc_64bit.go:235:9: inlining call to sysMapOS
runtime/mspanset.go:218:24: inlining call to (*spanSetBlockAlloc).free
runtime/mspanset.go:218:24: inlining call to atomic.(*Uint32).Store
runtime/mspanset.go:264:25: inlining call to (*spanSetBlockAlloc).free
runtime/mspanset.go:264:25: inlining call to atomic.(*Uint32).Store
runtime/pinner.go:171:34: inlining call to (*mspan).newPinnerBits
runtime/pinner.go:315:35: inlining call to (*mspan).newPinnerBits
runtime/print.go:252:14: inlining call to printpointer
runtime/proc.go:1276:16: inlining call to casGToWaiting
runtime/proc.go:1452:18: ... argument does not escape
runtime/proc.go:1452:18: inlining call to (*muintptr).set
runtime/proc.go:1452:18: inlining call to (*puintptr).set
runtime/proc.go:1452:18: inlining call to (*randomOrder).reset
runtime/proc.go:1452:18: inlining call to assertLockHeld
runtime/proc.go:1452:18: inlining call to assertWorldStopped
runtime/proc.go:1452:18: inlining call to atomicstorep
runtime/proc.go:1452:18: inlining call to gcd
runtime/proc.go:1452:18: inlining call to getLockRank
runtime/proc.go:1452:18: inlining call to lock
runtime/proc.go:1452:18: inlining call to lockWithRank
runtime/proc.go:1452:18: inlining call to mget
runtime/proc.go:1452:18: inlining call to muintptr.ptr
runtime/proc.go:1452:18: inlining call to nanotime
runtime/proc.go:1452:18: inlining call to noescape
runtime/proc.go:1452:18: inlining call to noescape
runtime/proc.go:1452:18: inlining call to procresize
runtime/proc.go:1452:18: inlining call to puintptr.ptr
runtime/proc.go:1452:18: inlining call to runqempty
runtime/proc.go:1452:18: inlining call to traceEnabled
runtime/proc.go:1452:18: inlining call to traceGoSched
runtime/proc.go:1452:18: inlining call to traceGomaxprocs
runtime/proc.go:1452:18: inlining call to unlock
runtime/proc.go:1452:18: inlining call to unlockWithRank
runtime/proc.go:1452:18: make([]*p, nprocs) escapes to heap
runtime/proc.go:1452:18: make([]uint32, maskWords) escapes to heap
runtime/proc.go:1452:18: new(p) escapes to heap
runtime/proc.go:4492:19: inlining call to (*gcControllerState).addScannableStack
runtime/proc.go:4492:19: inlining call to (*goroutineProfileStateHolder).Store
runtime/proc.go:4492:19: inlining call to acquirem
runtime/proc.go:4492:19: inlining call to alignUp
runtime/proc.go:4492:19: inlining call to atomic.(*Int32).Add
runtime/proc.go:4492:19: inlining call to atomic.(*Uint32).Load
runtime/proc.go:4492:19: inlining call to atomic.(*Uint32).Store
runtime/proc.go:4492:19: inlining call to atomic.(*Uint64).Add
runtime/proc.go:4492:19: inlining call to fastrand
runtime/proc.go:4492:19: inlining call to newproc1
runtime/proc.go:4492:19: inlining call to puintptr.ptr
runtime/proc.go:4492:19: inlining call to readgstatus
runtime/proc.go:4492:19: inlining call to releasem
runtime/proc.go:4492:19: inlining call to traceEnabled
runtime/stack.go:930:12: inlining call to (*unwinder).init
runtime/symtab.go:820:25: inlining call to (*moduledata).funcName
runtime/symtab.go:980:25: inlining call to (*moduledata).funcName
runtime/traceback.go:1114:8: inlining call to (*unwinder).init

Interesting. My best guess is that the additional PGO inlines made a caller of NewMsgUnknownSymbol too expensive to inline into. The "hot-budget" logs only inlines explicitly allowed by PGO. There could be some downstream effects on non-PGO inlines [1].

The difference results don't seem to support your guess, it shows what happened after b56645a87b28840a180d64077877cb46570b4176, gitee.com/u-language/u-language/ucom/check3/check3.go:45:39: inlining call to errcode.NewMsgUnknownSymbol but I don`t confirm this come from the performance regression line, because https://gitee.com/u-language/u-language/blob/pgo2024-8-9/ucom/check3/check3.go#L201 call same funcname in caller(checkCall).