April 2023 - Boost-users - lists.stage.boost.cppalliance.org

b2 error in nvidia enviroment
by Cho-Kuen Ng 26 Apr '23

26 Apr '23

I tried to build under nvidia environment and got the following b2 error: ../boost_1_73_0/tools/build/share/boost-build/src/build/toolset.jam:44: in toolset.using ERROR: rule "nvc.init" unknown in module "toolset". I put the following line in the site-config file using nvc : : CC ; What compiler toolset should I use under nvidia? Thanks.

2 1

Re: [Boost-users] errors in multiprecision
by John Maddock 21 Apr '23

21 Apr '23

On 20/04/2023 18:28, Gero Peterhoff via Boost-users wrote: > Am 20.04.23 um 18:38 schrieb John Maddock via Boost-users: >> On 20/04/2023 16:58, Gero Peterhoff via Boost-users wrote: >>> Hi John, >>> some functions do not work with multiprecision-integer-types, e.g. >>> >>> using T = boost::multiprecision::int256_t; >>> using ::std::isinf; >>> >>> const auto x = T{5}; >>> const auto y = isinf(x); // ct error >>> std::cout << y << std::endl; >>> >>> gives a CT error. >>> >> Correct. Those functions are pretty meaningless in the context of >> integer types, and no one has ever asked for them. >> >> John. > > > Ok, herewith I ask for it then :-) Finally this is relevant in the > template context or a functions with std- AND multiprecision-types > should/must work. I'll add it to the list, but be aware that this is not always supported for regular integer types: https://godbolt.org/#g:!((g:!((g:!((h:codeEditor,i:(filename:'1',fontScale:14,fontUsePx:'0',j:1,lang:c%2B%2B,selection:(endColumn:2,endLineNumber:6,positionColumn:2,positionLineNumber:6,selectionStartColumn:2,selectionStartLineNumber:6,startColumn:2,startLineNumber:6),source:'%23include+%3Ccmath%3E%0A%0Aint+main(int+argc)%0A%7B%0A++++return+std::isinf(argc)%3B%0A%7D%0A'),l:'5',n:'0',o:'C%2B%2B+source+%231',t:'0')),k:33.333333333333336,l:'4',n:'0',o:'',s:0,t:'0'),(g:!((h:compiler,i:(compiler:vcpp_v19_33_x64,deviceViewOpen:'1',filters:(b:'0',binary:'0',binaryObject:'1',commentOnly:'0',demangle:'0',directives:'0',execute:'0',intel:'0',libraryCode:'0',trim:'1'),flagsViewOpen:'1',fontScale:15,fontUsePx:'0',j:1,lang:c%2B%2B,libs:!(),options:'-O2',selection:(endColumn:1,endLineNumber:1,positionColumn:1,positionLineNumber:1,selectionStartColumn:1,selectionStartLineNumber:1,startColumn:1,startLineNumber:1),source:1),l:'5',n:'0',o:'+x64+msvc+v19.33+(Editor+%231)',t:'0')),k:33.333333333333336,l:'4',n:'0',o:'',s:0,t:'0'),(g:!((h:output,i:(compilerName:'x64+msvc+v19.latest',editorid:1,fontScale:14,fontUsePx:'0',j:1,wrap:'1'),l:'5',n:'0',o:'Output+of+x64+msvc+v19.33+(Compiler+%231)',t:'0')),k:33.33333333333333,l:'4',n:'0',o:'',s:0,t:'0')),l:'2',n:'0',o:'',t:'0')),version:4 John.

2 3

Re: [Boost-users] errors in multiprecision
by Gero Peterhoff 20 Apr '23

20 Apr '23

Am 20.04.23 um 18:38 schrieb John Maddock via Boost-users: > On 20/04/2023 16:58, Gero Peterhoff via Boost-users wrote: >> Hi John, >> some functions do not work with multiprecision-integer-types, e.g. >> >> using T = boost::multiprecision::int256_t; >> using ::std::isinf; >> >> const auto x = T{5}; >> const auto y = isinf(x); // ct error >> std::cout << y << std::endl; >> >> gives a CT error. >> > Correct. Those functions are pretty meaningless in the context of integer types, and no one has ever asked for them. > > John. Ok, herewith I ask for it then :-) Finally this is relevant in the template context or a functions with std- AND multiprecision-types should/must work. thx Gero

2 5

Re: [Boost-users] errors in multiprecision
by John Maddock 20 Apr '23

20 Apr '23

On 20/04/2023 16:58, Gero Peterhoff via Boost-users wrote: > Hi John, > some functions do not work with multiprecision-integer-types, e.g. > > using T = boost::multiprecision::int256_t; > using ::std::isinf; > > const auto x = T{5}; > const auto y = isinf(x); // ct error > std::cout << y << std::endl; > > gives a CT error. > Correct. Those functions are pretty meaningless in the context of integer types, and no one has ever asked for them. John.

1 0

errors in multiprecision
by Gero Peterhoff 20 Apr '23

20 Apr '23

Hi John, some functions do not work with multiprecision-integer-types, e.g. using T = boost::multiprecision::int256_t; using ::std::isinf; const auto x = T{5}; const auto y = isinf(x); // ct error std::cout << y << std::endl; gives a CT error. cu Gero

5 5

Re: [Boost-users] Much slower compilation time with boost-1.81 than boost-1.80 + much bigger .text section
by Niall Douglas 18 Apr '23

18 Apr '23

On 17/04/2023 14:33, Dominique Pellé via Boost-users wrote: > Hi > > 1) Compiling the attached "brands.cpp" file which uses > boost::container::flat_set is very slow with clang when using > boost-1.81.0. It was much faster with boost-1.80.0 somehow. Thanks for reporting this. Peter Dimov looked into this and found that the cause is recently added use of __forceinline. A suitable workaround until the next Boost release is -DBOOST_CONTAINER_DISABLE_FORCEINLINE I can't speak for Ion, but I'd assume it'll be fixed next Boost release. Once again thanks for reporting this, it wasn't noticed by the usual release testing. Niall

2 1

Much slower compilation time with boost-1.81 than boost-1.80 + much bigger .text section
by Dominique Pellé 18 Apr '23

18 Apr '23

Hi 1) Compiling the attached "brands.cpp" file which uses boost::container::flat_set is very slow with clang when using boost-1.81.0. It was much faster with boost-1.80.0 somehow. 2) Furthermore, the size of the .text section is much bigger when using boost-1.81.0 than with boost-1.80.0 when using clang. The source file is a simplification of an actual file in my project. Note that the code is in anonymous namespace (removing the anonymous namespace makes the compilation time much faster, presumably because there are more optimizations in the anonymous namespace). Benchmarks with clang-16 + boost-1.81.0 (on Linux Ubuntu-22.04): ``` $ time clang++-16 -isystem boost-1.81 -O2 -DNDEBUG \ -std=c++14 -o brands.o -c brands.cpp real 1m46.638s user 1m46.090s sys 0m0.520s $ size brands.o text data bss dec hex filename 2403932 20032 48 2424012 24fccc brands.o ``` Compare with the older boost-1.80.0: ``` $ time clang++-16 -isystem boost-1.80 -O2 -DNDEBUG \ -std=c++14 -o brands.o -c brands.cpp real 0m2.163s user 0m2.122s sys 0m0.041s $ size brands.o text data bss dec hex filename 101280 20032 48 121360 1da10 brands.o ``` In other words: - file is 81 times slower to compile with boost-1.81 than with boost-1.80 - the .text section is 23 times bigger with boost-1.81 than with boost-1.80 This may be a clang compiler issue, but since the performance regression is introduced from boost-1.80.0 to boost-1.81.0 I'm, reporting it here (I may report it to clang later depending on how the discussion here goes). It's even worse with older clang compilers. The table below shows the compilation time of brands.cpp file and size of the .text section for with various compilers using boost-1.80 and boost-1.81: ``` compiler boost compilation time size .text section -------- ----- ---------------- ------------------ clang-16 1.81 106.638 sec 2,403,932 bytes clang-16 1.80 2.164 sec 101,280 bytes clang-15 1.81 137.508 sec 2,861,332 bytes clang-15 1.80 2.527 sec 107,944 bytes clang-14 1.81 153.504 sec 3,033,487 bytes clang-14 1.80 2.294 sec 111,599 bytes clang-5.0 1.81 276.031 sec 2,720,187 bytes clang-5.0 1.80 2.328 sec 126,811 bytes gcc-12.1 1.81 2.821 sec 139,833 bytes gcc-12.1 1.80 2.505 sec 140,217 bytes gcc-7.5 1.81 1.943 sec 121,497 bytes gcc-7.5 1.80 1.707 sec 122,050 bytes ``` Regards Dominique PS: full versions of compiler used in above table (--version option): - Ubuntu clang version 16.0.2 (++20230414073413+b5aa566a7e53-1~exp1~20230414073428.73) - Ubuntu clang version 15.0.7 - Ubuntu clang version 14.0.0-1ubuntu1 - clang version 5.0.2 (tags/RELEASE_502/final) - gcc-12 (Ubuntu 12.1.0-2ubuntu1~22.04) 12.1.0 - gcc-7 (Ubuntu 7.5.0-6ubuntu2) 7.5.0

1 0

Re: [Boost-users] Boost configuration errors and problems
by Gavin Lambert 17 Apr '23

17 Apr '23

On 18/04/2023 02:32, Klebsch, Mario via wrote: > Since the vast majority of boost are header-only libraries, my usual > advice is no NOT use precompiled libraries of boost but to use boost > as a kind of source code library. This way, the source code is always > compiled with the same compiler flags as the rest of your code. > > I usually have to include > .../libs/system/src/error_code.cpp > .../libs/thread/src/pthread/* (or .../libs/thread/src/win32/*) > .../libs/chrono/src/* > .../libs/filesystem/src/* > In my Makefile, CMakeLists.txt or .vxcproj. Note that while doing this sort of thing is a valid strategy for a single application, it will get you into trouble if you split your app into separate EXE and DLL files and want to have any Boost types in the interfaces between them.

5 5

Re: [Boost-users] Boost configuration errors and problems
by Klebsch, Mario 17 Apr '23

17 Apr '23

Hello, > No suitable build variant has been found. > The following variants have been tried and rejected: > * boost_filesystem.lib (shared, Boost_USE_STATIC_LIBS=ON) Since the vast majority of boost are header-only libraries, my usual advice is no NOT use precompiled libraries of boost but to use boost as a kind of source code library. This way, the source code is always compiled with the same compiler flags as the rest of your code. I usually have to include .../libs/system/src/error_code.cpp .../libs/thread/src/pthread/* (or .../libs/thread/src/win32/*) .../libs/chrono/src/* .../libs/filesystem/src/* In my Makefile, CMakeLists.txt or .vxcproj. On Windows I also have to #define BOOST_ALL_NO_LIB to prevent the magic from trying to pick up pre-built libraries at link time and sometimes BOOST_THREAD_BUILD_LIB or BOOST_ALL_STATIC_LINK are required. Klebsch Mario Funktion | R&D Tel: +49 (0) 531 38 701 718 Raum: Braunschweig, E20 Diese E-Mail und die an sie angehängten Dateien sind ausschließlich für Personen oder Institutionen bestimmt, deren Namen oben aufgeführt sind. Sie können Informationen enthalten, die durch das Berufsgeheimnis geschützt sind und deren Weitergabe strengstens untersagt ist. Jede elektronische Nachricht kann geändert werden. ACTIA lehnt jede Verantwortung für diese Nachricht ab. Der Inhalt dieser Nachricht stellt keine Verpflichtung seitens unseres Unternehmens dar. Wenn Sie kein Empfänger sind, weisen wir Sie darauf hin, dass das Lesen, Vervielfältigen oder Verteilen strengstens untersagt ist. Wir bitten Sie daher, uns umgehend über diesen Brief zu informieren und diese Nachricht sowie eventuell beigefügte Unterlagen aus Ihrem Postfach zu löschen. Danke. This e-mail and the files attached to it are intended exclusively for persons or institutions whose names are listed above. They may contain information that is protected by professional secrecy and the disclosure of which is strictly prohibited. Any electronic message can be modified. ACTIA declines all responsibility for this message. The content of this message does not represent a commitment on the part of our company. If you are not a recipient, we would like to point out that reading, copying or distribution is strictly prohibited. We therefore ask you to inform us immediately about this letter and to delete this message and any accompanying documents from your mailbox. Thank you.

7 14

Re: [Boost-users] Precompiled Windows ARM/ARM64 binaries
by Tom Kent 14 Apr '23

14 Apr '23

On Fri, Apr 14, 2023 at 12:39 PM Stephan Brunner via Boost-users < boost-users(a)lists.boost.org> wrote: > Hi, > > are there any plans to ship precompiled windows binaries for ARM and/or > ARM64 at some time on SourceForge [1]? > Currently only a x86 and a x64 build is available there. > If not, I'll look into building my own version for those architectures. > > Thanks a lot! > > [1] https://sourceforge.net/projects/boost/files/boost-binaries/1.81.0/ > > No, I don't have a good windows arm/arm64 machine to test on, so don't want to make binaries I can't at least run a smoke test on. Beyond that, it would make the build process a lot more complex to run on multiple machines. The nice thing about i368 and amd64 is that you can run them both on the same windows install! Tom

1 0