Le 13/02/2019 à 20:53, clipka a écrit :
> But as I presume this is using the Windows scheduler, I expect that
> genuine Linux systems may behave differently, and I'm also interested in
> other platforms (Mac, maybe BSD if some of you folks are using that, or
> actually any system you can get your hands on.)
> 
> I have also reason to believe that results may differ between compilers.
> Using g++ 5.4.0 here.

$ lsb_release -a
No LSB modules are available.
Distributor ID:	Ubuntu
Description:	Ubuntu 18.10
Release:	18.10
Codename:	cosmic

====

For gcc:

$ time ./a.out
Done.

real	0m1,096s
user	0m0,000s
sys	0m0,015s

$ g++ --version
g++ (Ubuntu 8.2.0-7ubuntu1) 8.2.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.


====

For icpc:

$ time ./a.out
Done.

real	0m1,094s
user	0m0,000s
sys	0m0,015s

$ icpc --version
icpc (ICC) 19.0.1.144 20181018
Copyright (C) 1985-2018 Intel Corporation.  All rights reserved.


====

For clang:

$ time ./a.out
Done.

real	0m1,094s
user	0m0,015s
sys	0m0,000s

$ clang++ --version
clang version 7.0.0-3 (tags/RELEASE_700/final)
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

Post a reply to this message

Am 15.02.2019 um 08:40 schrieb ingo:
> compiling with the wrong compiler and wrong option creates no errors and
> a program that won't run but you don't notice because the test file test
> is not called but test (check file types and compare values) that is
> already present on the system...

Oh yeah - that's a good one :)

Even better than trying to judge the run-time performance of certain 
code under optimization, and typing `g++ -o3 -ffast-math -o test 
test.cpp` instead of `g++ -O3 -ffast-math -o test test.cpp`...

> But, the proper results are in:
> 
> FreeBSD Clang 6.0
> Delay 1
> 0.000u 0.005s 0:01.07 0.0%      0+0k 2+0io 0pf+0w
> 
> Delay 10
> 0.000u 0.006s 0:10.64 0.0%      0+0k 6+0io 15pf+0w
> 
> Delay 100
> 0.000u 0.005s 1:46.24 0.0%      0+0k 2+0io 0pf+0w

That's 1.07 ms, 10.64 ms and 106.24 ms, respectively, which makes a lot 
more sense :)

> Mingwin g++.exe (Rev1, Built by MSYS2 project) 8.2.1 20181207
> Delay 1
> real    0m15.822s
> user    0m0.000s
> sys     0m0.015s
> 
> Delay 10
> real    0m15.759s
> user    0m0.000s
> sys     0m0.031s
> 
> Delay 100
> real    1m49.566s
> user    0m0.015s
> sys     0m0.031s
> 
> ingo

That, too, makes sense, and indicates that MinGW has a minimum wait of 
ca. 15 ms (on average at any rate). Which is fine, as C++11 allows 
implementations to wait longer than requested (but not less).

Post a reply to this message

Le 13/02/2019 à 20:53, clipka a écrit :
> Hi folks,
> 
> can a few of you folks please run the following little program on your 
> computers and report the time it takes to run?
> 
> --------------------------------------------------
> #include <chrono>
> #include <thread>
> #include <iostream>
> 
> inline void Delay(unsigned int msec)
> {
>      std::this_thread::sleep_for(std::chrono::milliseconds(msec));
> }
> 
> int main()
> {
>      int count =  1000;
>      for (int i = 0; i < count; ++i)
>      {
>          Delay(1);
>      }
>      std::cout << "Done." << std::endl;
> }
> --------------------------------------------------
> 
> On Windows Subsystem For Linux I see results like the following:
> 
> real    0m1.782s
> user    0m0.000s
> sys     0m0.000s
> 
> But as I presume this is using the Windows scheduler, I expect that 
> genuine Linux systems may behave differently, and I'm also interested in 
> other platforms (Mac, maybe BSD if some of you folks are using that, or 
> actually any system you can get your hands on.)
> 
> I have also reason to believe that results may differ between compilers. 
> Using g++ 5.4.0 here.

my little contribution on Mac Intel i5 - 2,7 Ghz.

g++ -v :
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr 
--with-gxx-include-dir=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.10.sdk/usr/include/c++/4.2.1
Apple LLVM version 6.0 (clang-600.0.57) (based on LLVM 3.5svn)
Target: x86_64-apple-darwin13.4.0
Thread model: posix

uname -v :
Darwin Kernel Version 13.4.0: Mon Jan 11 18:17:34 PST 2016; 
root:xnu-2422.115.15~1/RELEASE_X86_64


result :

	real	0m1.239s
	user	0m0.009s
	sys	0m0.019s






-- 
Kurtz le pirate
Compagnie de la Banquise

Post a reply to this message