gprof 不为需要合理时间执行的程序产生任何输出
Posted
技术标签:
【中文标题】gprof 不为需要合理时间执行的程序产生任何输出【英文标题】：gprof producing no output for a program that takes reasonable time to execute 【发布时间】：2021-07-18 07:14:07 【问题描述】：我知道以前有人问过类似的问题，但是：
-
关于gprof produces empty output，我在 Windows 10 上使用来自 MSYS2 的 GCC 10.2.0，即它是一个 MingW64 发行版。我还在链接器参数中添加了
-no-pie，但仍然没有结果。 GNU Prof 的版本是 2.36.1
关于gprof gives no output，我正在一个明确需要时间执行的程序上执行此操作。
该程序是来自this tutorial的代码，复制如下：
//test_gprof.c
#include<stdio.h>
void new_func1(void)
printf("\n Inside new_func1()\n");
int i = 0;
for(;i<0xffffffee;i++);
return;
void func1(void)
printf("\n Inside func1 \n");
int i = 0;
for(;i<0xffffffff;i++);
new_func1();
return;
static void func2(void)
printf("\n Inside func2 \n");
int i = 0;
for(;i<0xffffffaa;i++);
return;
int main(void)
printf("\n Inside main()\n");
int i = 0;
for(;i<0xffffff;i++);
func1();
func2();
return 0;
我正在使用 Code::Blocks 进行构建。我已经尝试过 MSYS2 的 32 位和 64 位版本的 MingW。我已将所有内容升级到可用软件包的最新版本。
Code::Blocks 构建日志的输出：
g++.exe -pg -c C:\Users\david\Documents\GameDev\CPPTESTS\main.cpp -o obj\Debug\CPPTESTS\main.o
g++.exe -o CPPTESTS-d.exe obj\Debug\CPPTESTS\main.o -pg -lgmon -no-pie
程序编译执行成功。执行后，会生成一个 gmon.out 文件。执行 gprof 来解释这个文件：
gprof CPPTests-d.exe gmon.out > gprofoutput.txt
在gprofoutput.txt 中产生以下内容：
Flat profile:
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls Ts/call Ts/call name
% the percentage of the total running time of the
time program used by this function.
cumulative a running sum of the number of seconds accounted
seconds for by this function and those listed above it.
self the number of seconds accounted for by this
seconds function alone. This is the major sort for this
listing.
calls the number of times this function was invoked, if
this function is profiled, else blank.
self the average number of milliseconds spent in this
ms/call function per call, if this function is profiled,
else blank.
total the average number of milliseconds spent in this
ms/call function and its descendents per call, if this
function is profiled, else blank.
name the name of the function. This is the minor sort
for this listing. The index shows the location of
the function in the gprof listing. If the index is
in parenthesis it shows where it would appear in
the gprof listing if it were to be printed.
Copyright (C) 2012-2021 Free Software Foundation, Inc.
Copying and distribution of this file, with or without modification,
are permitted in any medium without royalty provided the copyright
notice and this notice are preserved.
Call graph (explanation follows)
granularity: each sample hit covers 4 byte(s) no time propagated
index % time self children called name
This table describes the call tree of the program, and was sorted by
the total amount of time spent in each function and its children.
Each entry in this table consists of several lines. The line with the
index number at the left hand margin lists the current function.
The lines above it list the functions that called this function,
and the lines below it list the functions this one called.
This line lists:
index A unique number given to each element of the table.
Index numbers are sorted numerically.
The index number is printed next to every function name so
it is easier to look up where the function is in the table.
% time This is the percentage of the `total' time that was spent
in this function and its children. Note that due to
different viewpoints, functions excluded by options, etc,
these numbers will NOT add up to 100%.
self This is the total amount of time spent in this function.
children This is the total amount of time propagated into this
function by its children.
called This is the number of times the function was called.
If the function called itself recursively, the number
only includes non-recursive calls, and is followed by
a `+' and the number of recursive calls.
name The name of the current function. The index number is
printed after it. If the function is a member of a
cycle, the cycle number is printed between the
function's name and the index number.
For the function's parents, the fields have the following meanings:
self This is the amount of time that was propagated directly
from the function into this parent.
children This is the amount of time that was propagated from
the function's children into this parent.
called This is the number of times this parent called the
function `/' the total number of times the function
was called. Recursive calls to the function are not
included in the number after the `/'.
name This is the name of the parent. The parent's index
number is printed after it. If the parent is a
member of a cycle, the cycle number is printed between
the name and the index number.
If the parents of the function cannot be determined, the word
`<spontaneous>' is printed in the `name' field, and all the other
fields are blank.
For the function's children, the fields have the following meanings:
self This is the amount of time that was propagated directly
from the child into the function.
children This is the amount of time that was propagated from the
child's children to the function.
called This is the number of times the function called
this child `/' the total number of times the child
was called. Recursive calls by the child are not
listed in the number after the `/'.
name This is the name of the child. The child's index
number is printed after it. If the child is a
member of a cycle, the cycle number is printed
between the name and the index number.
If there are any cycles (circles) in the call graph, there is an
entry for the cycle-as-a-whole. This entry shows who called the
cycle (as parents) and the members of the cycle (as children.)
The `+' recursive calls entry shows the number of function calls that
were internal to the cycle, and the calls entry for each member shows,
for that member, how many times it was called from other members of
the cycle.
Copyright (C) 2012-2021 Free Software Foundation, Inc.
Copying and distribution of this file, with or without modification,
are permitted in any medium without royalty provided the copyright
notice and this notice are preserved.
Index by function name
你可以看到它完全是空的。此外，如果我在十六进制编辑器中打开gmon.out，里面大部分是空白的，这是第一位：
F0 14 C5 00 18 8A C5 00 B4 3A 00 00 79 18 05 00 64 00 00 00 96 1A C5 00 A7 16 C5 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 05 00 FE 00 5D 00 E2 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 07 00 C6 01 A5 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 06 00 E5 00 77 00 F5 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
然后结尾如下：
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 E4 15 C5 00 6C 89 C5 00 01 00 00 00 14 16 C5 00 6C 89 C5 00 01 00 00 00 2C 16 C5 00 DB 15 C5 00 01 00 00 00 48 16 C5 00 6C 89 C5 00 01 00 00 00 84 16 C5 00 6C 89 C5 00 01 00 00 00 A0 16 C5 00 0A 16 C5 00 01 00 00 00 A4 16 C5 00 3C 16 C5 00 01 00 00 00
剩下的是几千字节的00。
我已经尝试了所有我可以在谷歌上找到的关于让它工作的方法。最初我在项目的实际代码上执行它时遇到了麻烦，但我使用了上面的代码，因为我似乎做错了什么，我无法弄清楚。
任何帮助将不胜感激。
编辑
我设法在 gmon.out 中生成了可以用 gprof 正确解析的结果，并为我的测试程序中的所有函数提供了时间。不做任何更改，甚至不关闭 MSYS2 终端（我放弃了 Code::Blocks 以确保它不是问题），然后我使用相同的命令重新编译程序，再次执行它，然后返回一个空结果。
我不清楚它是在编译过程中发生的（难以置信，它是使用终端历史中的最后一个命令编译的），还是在我执行它时在我的机器上发生了什么。或者如果 gprof 翻译 gmon.out 文件有问题。
编辑 2
我在从未用于开发的另一台机器（类似的 Windows 10 Surface Pro）上执行了该程序。它产生的 gmon.out 在通过 gprof 解析时似乎也没有结果。
编辑 3
我已经厌倦了 gcc 和 g++ 的相同结果。我尝试过单独的编译和链接步骤，或者两者都做。我也尝试过链接到-lgmon 或没有。解析 gprof 时，All 给出相同的空结果。
自从那次随机出现后，我再也没有得到结果。
编辑 4
我用-g 和-pg 编译了程序，然后通过VTune 运行它。我可以在热点分析中看到正在调用与分析器相关的函数调用。
这是整个被调用者堆栈。您可以看到与配置文件相关的函数调用profthr_func 和get_thrpc 被调用。
Callees CPU Time: Total CPU Time: Self
BaseThreadInitThunk 100.0% 0s
_tmainCRTStartup 99.1% 0s
main 99.1% 0.016s
func1 65.8% 20.521s
new_func1 32.8% 20.456s
func2 33.2% 20.695s
profthr_func 0.6% 0.009s
WaitForSingleObject 0.4% 0.231s
get_thrpc 0.2% 0s
SwitchToThread 0.2% 0.116s
TlsGetValue 0.0% 0.002s
GetLastError 0.0% 0.001s
SwitchToThread 0.0% 0.001s
GetLastError 0.0% 0.001s
[Outside any known module] 0.0% 0.001s
func@0x4b308730 0.0% 0.000s
TlsGetValue 0.0% 0.000s
get_thrpc 0.4% 0s
GetThreadContext 0.3% 0.214s
GetLastError 0.0% 0.006s
TlsGetValue 0.0% 0.000s
[Outside any known module] 0.0% 0.002s
SuspendThread 0.0% 0.001s
[Unknown stack frame(s)] 0.0% 0s
profthr_func 0.0% 0s
WaitForSingleObject 0.0% 0.000s
然后是所有功能及其来源等：
Function CPU Time: Total CPU Time: Self Module Function (Full) Source File Start Address
BaseThreadInitThunk 100.0% 0s kernel32.dll BaseThreadInitThunk [Unknown] 0x6b81fa10
func@0x4b2e7a03 100.0% 0s ntdll.dll func@0x4b2e7a03 [Unknown] 0x4b2e7a03
func@0x4b2e7a1f 100.0% 0s ntdll.dll func@0x4b2e7a1f [Unknown] 0x4b2e7a1f
_tmainCRTStartup 99.1% 0s CPPTESTS-d.exe _tmainCRTStartup crtexe.c 0x401170
main 99.1% 0.016s CPPTESTS-d.exe main main.cpp 0x40165c
func1 65.8% 20.521s CPPTESTS-d.exe func1(void) main.cpp 0x4015fd
func2 33.2% 20.695s CPPTESTS-d.exe func2 main.cpp 0x40162f
new_func1 32.8% 20.456s CPPTESTS-d.exe new_func1(void) main.cpp 0x4015d0
profthr_func 0.6% 0.009s CPPTESTS-d.exe profthr_func profil.c 0x408620
get_thrpc 0.5% 0s CPPTESTS-d.exe get_thrpc profil.c 0x4085c0
WaitForSingleObject 0.4% 0.231s KernelBase.dll WaitForSingleObject [Unknown] 0x10110440
GetThreadContext 0.3% 0.214s KernelBase.dll GetThreadContext [Unknown] 0x101b61a0
SwitchToThread 0.2% 0.116s KernelBase.dll SwitchToThread [Unknown] 0x100f5b40
GetLastError 0.0% 0.007s kernel32.dll GetLastError [Unknown] 0x6b81e010
[Outside any known module] 0.0% 0.002s [Unknown] [Outside any known module] [Unknown] 0
TlsGetValue 0.0% 0.002s KernelBase.dll TlsGetValue [Unknown] 0x101193a0
SuspendThread 0.0% 0.001s kernel32.dll SuspendThread [Unknown] 0x6b834e90
SwitchToThread 0.0% 0.001s kernel32.dll SwitchToThread [Unknown] 0x6b817cc0
GetLastError 0.0% 0.001s KernelBase.dll GetLastError [Unknown] 0x10110580
TlsGetValue 0.0% 0.000s kernel32.dll TlsGetValue [Unknown] 0x6b81df20
func@0x4b308730 0.0% 0.000s ntdll.dll func@0x4b308730 [Unknown] 0x4b308730
[Unknown stack frame(s)] 0.0% 0s [Unknown] [Unknown stack frame(s)] [Unknown] 0
我不确定这些信息是否有助于解决我的问题。 gmon.out 文件再次通过 gprof 解析时没有任何信息。
我读过this answer to a question about how gprof works，所以我希望对main、func1、func2 和new_func1 的调用会包含某种仪器调用？
我运行了 VTune 并检查了 Threading 选项，并注意到 gprof 似乎 通过一个新线程来完成它，这与我上面的理解不同。但是，它并没有解释为什么我曾经得到一些输出并且无法重新创建它。
但是，在 Vtune 中，我没有看到对 mcount 的调用，我认为这是 gprof 的工作原理。它确实向我展示了来自 -pg 标志的分析 stuff 存在的点点滴滴（如您在上面的调用堆栈和函数列表中所见）。
编辑 5
我已经上传了我的compiled program，它是gmon.out，以防它帮助任何人帮助回答问题。
编辑 6
我重新安装了 MSYS2，希望能解决问题。没有。
编辑 7
我已尝试与-static -static-libgcc -static-libstdc++ 链接。没有效果。
编辑 8
我已经使用-d 标志执行了 gprof。 Debug output can be found here。但这对我来说大多只是胡言乱语。
编辑 9
我直接在 MSYS2 外部安装了Mingw-Builds，并在 i686 和 86_64 版本中测试了 8.1.0 版、Dwarf 和 SJLJ。用这个版本的 G++ 编译和链接后，会生成一个 gmon.out 文件，但是 gprof 然后给出以下错误（或者是数字的变化）：
BFD: Dwarf Error: Could not find abbrev number 108.
但生成的输出有效。我还检查了一个 8.1.0 编译的程序 gmon.out 可以被较新的 MSYS2 gprof 解析而没有问题。所以看来问题出在较新版本的 Mingw64 上。
Gprof 报告版本 2.30，G++ 报告版本 8.1.0。
编辑 10
我尝试在干净的 Windows 7 Virtual Box 机器上编译和运行 gprof 并获得相同的结果。所以我不是我的操作系统（Windows 10）或我的特定计算机。
这里似乎没有回答这个问题。如果我弄清楚了，我会回答这个问题。在那之前，我已经在MSYS2 tracker 和Mingw64 tracker 上创建了问题。
编辑 11
我现在一直在编写以下代码：
#include <iostream>
bool is_prime(const int& number)
if(number == 0 || number == 1)
return false;
else
for(int i = 2; i <= number / 2; ++i)
if(number % i == 0)
return false;
return true;
int main()
int low = 0;
int high = 300000;
while(low < high)
if(is_prime(low))
std::cout << low << ", ";
++low;
std::cout << std::endl;
return 0;
使用与之前 g++ -g -pg -O0 main.cpp -o CPPTESTS-d.exe 相同的编译器标志
objdump -d CPPTESTS-d.exe 为我提供了来自 main 和 is_prime 部分的以下内容：
004015d0 <__Z8is_primeRKi>:
4015d0: 55 push %ebp
4015d1: 89 e5 mov %esp,%ebp
4015d3: 83 ec 10 sub $0x10,%esp
4015d6: e8 a5 15 00 00 call 402b80 <_mcount>
4015db: 8b 45 08 mov 0x8(%ebp),%eax
4015de: 8b 00 mov (%eax),%eax
4015e0: 85 c0 test %eax,%eax
4015e2: 74 0a je 4015ee <__Z8is_primeRKi+0x1e>
4015e4: 8b 45 08 mov 0x8(%ebp),%eax
4015e7: 8b 00 mov (%eax),%eax
4015e9: 83 f8 01 cmp $0x1,%eax
4015ec: 75 07 jne 4015f5 <__Z8is_primeRKi+0x25>
4015ee: b8 00 00 00 00 mov $0x0,%eax
4015f3: eb 3b jmp 401630 <__Z8is_primeRKi+0x60>
4015f5: c7 45 fc 02 00 00 00 movl $0x2,-0x4(%ebp)
4015fc: 8b 45 08 mov 0x8(%ebp),%eax
4015ff: 8b 00 mov (%eax),%eax
401601: 89 c2 mov %eax,%edx
401603: c1 ea 1f shr $0x1f,%edx
401606: 01 d0 add %edx,%eax
401608: d1 f8 sar %eax
40160a: 39 45 fc cmp %eax,-0x4(%ebp)
40160d: 7f 1c jg 40162b <__Z8is_primeRKi+0x5b>
40160f: 8b 45 08 mov 0x8(%ebp),%eax
401612: 8b 00 mov (%eax),%eax
401614: 99 cltd
401615: f7 7d fc idivl -0x4(%ebp)
401618: 89 d0 mov %edx,%eax
40161a: 85 c0 test %eax,%eax
40161c: 75 07 jne 401625 <__Z8is_primeRKi+0x55>
40161e: b8 00 00 00 00 mov $0x0,%eax
401623: eb 0b jmp 401630 <__Z8is_primeRKi+0x60>
401625: 83 45 fc 01 addl $0x1,-0x4(%ebp)
401629: eb d1 jmp 4015fc <__Z8is_primeRKi+0x2c>
40162b: b8 01 00 00 00 mov $0x1,%eax
401630: c9 leave
401631: c3 ret
00401632 <_main>:
401632: 8d 4c 24 04 lea 0x4(%esp),%ecx
401636: 83 e4 f0 and $0xfffffff0,%esp
401639: ff 71 fc push -0x4(%ecx)
40163c: 55 push %ebp
40163d: 89 e5 mov %esp,%ebp
40163f: 51 push %ecx
401640: 83 ec 24 sub $0x24,%esp
401643: e8 38 15 00 00 call 402b80 <_mcount>
401648: e8 73 1a 00 00 call 4030c0 <__monstartup>
40164d: e8 be 01 00 00 call 401810 <___main>
401652: c7 45 f0 00 00 00 00 movl $0x0,-0x10(%ebp)
401659: c7 45 f4 e0 93 04 00 movl $0x493e0,-0xc(%ebp)
401660: 8b 45 f0 mov -0x10(%ebp),%eax
401663: 39 45 f4 cmp %eax,-0xc(%ebp)
401666: 7e 3d jle 4016a5 <__fu0___ZSt4cout+0x27>
401668: 8d 45 f0 lea -0x10(%ebp),%eax
40166b: 89 04 24 mov %eax,(%esp)
40166e: e8 5d ff ff ff call 4015d0 <__Z8is_primeRKi>
401673: 84 c0 test %al,%al
401675: 74 23 je 40169a <__fu0___ZSt4cout+0x1c>
401677: 8b 45 f0 mov -0x10(%ebp),%eax
40167a: 89 04 24 mov %eax,(%esp)
40167d: b9 .byte 0xb9
我可以看到mcount 在两者的开头都被调用了。所以它似乎编译的代码被正确检测，但在 gprof 中仍然没有输出。
【问题讨论】：
我刚刚使用了您的示例源代码，按照您的步骤操作，有无-no-pie，我得到了预期的结果，就像在您的链接教程中一样。所以很难说为什么它不适合你。
谢谢@ssbssa，请问您使用的是什么平台等？
还有 MSYS2 和 mingw-w64 gcc 10.2。
@ssbssa 您能否报告编译后的可执行文件的大小以及生成的 gmon.out 文件的大小/内容？
是让程序运行完成，还是强制关闭？如果异常终止，可能无法正确写入gmon.out。
【参考方案1】：
这似乎是新版本 MinGW 中的一个错误。我之前也遇到过同样的问题。我的 MinGW 是从 winlibs 下载的。当我尝试使用最新版本 9.0 时，gprof 的输出结果确实为空。我试过MinGW 9.0版和GCC 11、10、9版，结果都是空的。终于找到了mingw7.0.0-gcc10.2，尝试用这个版本编译，运行然后使用gprof，输出结果不为空，但是和Linux平台上使用gprof相比，结果有点奇怪
【讨论】：
以上是关于gprof 不为需要合理时间执行的程序产生任何输出的主要内容，如果未能解决你的问题，请参考以下文章