Add cmake options to include only specified modules in the library
Hi This option is not yet available when using cmake. You could try to change itpp/CMakeLists.txt file https://sourceforge.net/p/itpp/git/ci/master/tree/itpp/CMakeLists.txt by commenting the sources you don't want to be included, e.g., file ( GLOB ITPP_SRCS "base/*.cpp" "base/algebra/*.cpp" "base/bessel/*.cpp" "base/math/*.cpp" "comm/*.cpp" "fixed/*.cpp" "optim/*.cpp" # "protocol/*.cpp" # "signal/*.cpp" "srccode/*.cpp" "stat/*.cpp" ) regards Bogdan
Hi You are trying to use the old build system which is not longer maintained. Please use instead cmake: http://itpp.sourceforge.net/4.3.1/installation.html#inst_instr_cmake Bogdan
Hi Can you compile and link separately using BLAS, LAPACK and FFT libraries you found ? Also, as a first step, try to compile the library as static, but link with the shared versions of the external libs. Bogdan
Hi That message is a warning, BLAS library might not expose that specific function, this however does not impact the library functionality, probably only the processing time. Bogdan
Hi Heera Your message above says that there are compile time errors about missing dependences. So what are these errors specifically ? Bogdan
Update MKL support
Thanks for reporting this, I'll create a bug report
Hi This is a question you should address in MKL forums. One possible reason could be that IT++ is not using optimally the MKL API, but you should try to track down this problem to a specific call to be sure. regards Bogdan
Hi You are using the correct variabled for specifying the libraries, LAPACK_LIBRARIRES and BLAS_LIBRARIES. Try the following: - when you cross-compile use make VERBOSE=1 and check if the library is compiled using the correct BLAS/LAPACK libraries - make sure that when you compile using the static IT++ library you also add the shared libraries (if any) needed by IT++ regards Bogdan
It is itpp-4.3.1.tar.bz2 folder
Hi I did this on ubuntu 20 sudo apt install liblapack-dev libfftw3-dev then in IT++ source folder mkdir build && cd build cmake .. -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/itpp-install make -j make install
Hi I seems that FFTW and Lapack libs are not detected. Are you sure that they are installed ? Also, on which Ubuntu version this happens ? Bogdan
Hi itexports.h is generated based on itexports.h.cmake. Cmake replaces some placeholders found in itexports.h.cmake with their actual value, then renames the file. You can remove that include and define the missing variables as needed. Bogdan
Replace AMD Core Math Library (ACML) with AMD Optimizing CPU Libraries (AOCL)
Hi IT++ depends on several external libraries in order to speed up calculations: FFTW, Lapack, Blas. You can either find these libraries compiled for your mobile platform or try to compile IT++ without these libraries (probably some speed penality will occur). Besides external dependences, since IT++ library is written in C++ and uses cmake as compilation management system, it can be crosscompiled on other platforms (e.g. Android, iOS) using toolchain files (https://cmake.org/cmake/help/latest/manual/cmake-toolchains.7.html)....
I have found another a problem, when selecting only several items from the above archive (total size uncompressed 1.42 GB). This time 7zip application has the same bug: total size uncompressed is shown 1.42 GB, but actually the size is 780 MB. Any ideeas ?
I have still found a problem, when selecting only several items from the above archive (total size uncompressed 1.42 GB). This time 7zip application has the same bug: total size uncompressed is shown 1.42 GB, but actually the size is 780 MB. Any ideeas ?
The problem is solved after sorting the indices. Thank you @ipavlov for your help.
CArchiveExtractCallback *extractCallbackSpec = new CArchiveExtractCallback; CMyComPtr<IArchiveExtractCallback> extractCallback(extractCallbackSpec); UInt32 *indexes = (UInt32*)malloc(extractOptions.m_IndexesToExtract.count() * sizeof(UInt32)); inArchive->Extract(indexes, extractOptions.m_IndexesToExtract.count(), false, extractCallback); extractOptions.m_IndexesToExtract.count() has value 2883 indexes are not sorted however
It seems that extracted sizes around 1GB give wrong results, while size about 500 MB gives correct results 3.7z: 1.12 GB, extracted 1.42 GB, reported 5 GB 3_1.7z: 977.66 MB, extracted 1.19 GB, reported 2.5 GB 3_2.7z: 729.91 MB, extracted 974 MB, reported 2.61 GB 3_3.7z: 408.33 MB, extracted 490 MB, reported 489.22 MB The above results are obtained with my application using COM interface. 7z Windows application gives always correct results.
It seems that extracted sizes around 1GB give wrong results, while size about 500 MB gives correct results 3.7z: 1.12 GB, extracted 1.42 GB, reported 5 GB 3_1.7z: 977.66 MB, extracted 1.19 GB, reported 2.5 GB 3_2.7z: 729.91 MB, extracted 974 MB, reported 2.61 GB 3_3.7z: 408.33 MB, extracted 490 MB, reported 489.22 MB
It seems that extracted sizes around 1GB give wrong results, while size about 500 MB give correct results 3.7z: 1.12 GB, extracted 1.42 GB, reported 5 GB 3_1.7z: 977.66 MB, extracted 1.19 GB, reported 2.5 GB 3_2.7z: 729.91 MB, extracted 974 MB, reported 2.61 GB 3_3.7z: 408.33 MB, extracted 490 MB, reported 489.22 MB
It seems that extracted size around 1GB give wrong results, while size about 500 MB give correct results 3.7z: 1.12 GB, extracted 1.42 GB, reported 5 GB 3_1.7z: 977.66 MB, extracted 1.19 GB, reported 2.5 GB 3_2.7z: 729.91 MB, extracted 974 MB, reported 2.61 GB 3_3.7z: 408.33 MB, extracted 490 MB, reported 489.22 MB
correct size: 1.42 GB list command size: 1531915165 1207211174 2590 files, 293 folders extract command size: see my post above incorrect size: 5GB. This size is reported by SetTotal method of the callback COM object passed to IInArchive::Extract() method. SetCompleted method, that reports the extraction progress, gives results consistent with SetTotal method, i.e. from 0 to about 5GB
correct size: 1.42 GB list command size: 1531915165 1207211174 2590 files, 293 folders extract command size: see my post above incorrect size: 5GB. This size is reported by the SetTotal method of the callback COM object passed to IInArchive::Extract() method. SetCompleted method, that reports the extraction progress, gives results consistent with SetTotal method, i.e. from 0 to about 5GB
correct size: 1.42 GB list command size: 1531915165 1207211174 2590 files, 293 folders extract command size: see my post above incorrect size: 5GB. This size is reported by the SetTotal method of the callback COM object passed to IInArchive::Extract() method. SetCompleted method, that reports the extration progress, gives results consistent with SetTotal method, i.e. from 0 to about 5GB
correct size: 1.42 GB list command size: 1531915165 1207211174 2590 files, 293 folders extract command size: see my post above incorrect size: 5GB (this size is reported by the SetTotal method of the callback COM object passed to IInArchive::Extract() method)
correct size: 1.42 GB list command size: 1531915165 1207211174 2590 files, 293 folders extract command size: see my post above incorrect size: 5GB (this size is reported by the callback COM object passed to IInArchive::Extract() method)
Hi Are there instructions how to compile 7z.dll from the sources ? I might be able in this way to track down the above problem. thanks Bogdan
In my above comment the reported size is the one provided by SetTotal callback
C:\Program Files\7-Zip>7z.exe x -y C:/sandbox/3.7z -oC:/sandbox/temp -r 7-Zip 19.00 (x64) : Copyright (c) 1999-2018 Igor Pavlov : 2019-02-21 Scanning the drive for archives: 1 file, 1207225451 bytes (1152 MiB) Extracting archive: C:\sandbox\3.7z Path = C:\sandbox\3.7z Type = 7z Physical Size = 1207225451 Headers Size = 14277 Method = LZMA:24 BCJ Solid = + Blocks = 2 Everything is Ok Folders: 293 Files: 2590 Size: 1531915165 Compressed: 1207225451
How the logs can be enabled ? I am using the DLL provided by 7z installer.
How the logs can be enabled ? I am using the DLL provided by 7z installer.
Hi For some 7z archives the reproted total size (after decompression) is wrong, 5 GB instead of 1.42 GB. For other archives the decompressed size is correct. 7z application reports always the correct size. Is there an explanation for this ? thanks Bogdan
Hi Check cmake logs in order to learn why FindITPP.cmake failed. regards Bogdan
Hi This is a draft paper that explains the algorithms implemented in IT++ for indoor localisation. Check the bibliography for references: https://www.researchgate.net/profile/Bogdan_Cristea/publication/311695923_Hybrid_Multilateration_for_Indoor_Localization/links/58552c0908ae81995eb3f2f9/Hybrid-Multilateration-for-Indoor-Localization.pdf regards Bogdan
you need to set the library search path to the location where libitpp.so is installed: use either LD_LIBRARY_PATH or ldconfig application
Hi Here are the steps I have used: git clone ssh://<user>@git.code.sf.net/p/itpp/git itpp-git git clone https://github.com/google/googletest.git cmake .. -DGTEST_DIR=$HOME/projects/googletest/googletest Note that GTEST_DIR does not point to the sources root folder (when googletest is cloned with git) but goes one level lower where only googletest sources are. In order to run the tests use (from build folder) ./gtests/itpp_gtests
Hi gtest sources need to be downloaded separately in your gtests folder before you start the actual compilation. regards Bogdan
Hi The above change should be made in case IT++ compilation gives an error. If you have an error either you change random_dsfmt.h file or checkout with git the master branch. If no error nothing should be done. regards Bogdan
Hi I recommend to checkout the master branch and compile from there IT++. Here is the change I have made: --- a/itpp/base/random_dsfmt.h +++ b/itpp/base/random_dsfmt.h @@ -299,7 +299,7 @@ */ static void do_recursion(typename Context::w128_t *r, typename Context::w128_t *a, typename Context::w128_t *b, typename Context::w128_t *lung) { #if defined(__SSE2__) - const unsigned int SSE2_SHUFF = 0x1bU; +#define SSE2_SHUFF 0x1bU __m128i x = a->si; __m128i z = _mm_slli_epi64(x, SL1);
Hi I have solved recently a similar issue that generated errors when IT++ is compiled on Ubuntu Bionic. You can find the relevant change here: https://sourceforge.net/p/itpp/git/ci/bb5c7e95f40e8fdb5c3f3d01a84bcbaf76f3676d/tree/itpp/base/random_dsfmt.h?diff=d425a09330ced7aa8192d695df9d38876df6c76d
Hi I have provided a fix into the master branch. regards Bogdan
corrected compilation on Ubuntu Bionic
Hi IT++ is released under GPL licence and there were some discussions to adopt LGPL. If adopted, probably this will remain the only option for a commercial version. regards Bogdam
Merge commit 'f55b245697f91e9fafb22147375d631b1058b5f6'
Merge commit '5380e907ef29344ce0b318c96f96419be3c0cb11'
Hi From command line use cmake .. -DITPP_EXCEPTIONS=on or in cmake-gui define a new variable ITPP_EXCEPTIONS of type bool and init with true. This should at least put in config.h a #define ITPP_EXCEPTIONS regards Bogdan
Hi Sorry, my mistake, it has been some time since I haven't used SISO modules. There...
Hi This is done internally using the puncturing matrix. regards Bogdan
For tests on Windows this service https://www.appveyor.com is free for open-source...
Hi I have just received an e-mail from someone interested in using IT++ algorithms...
GPL->LGPL licence change, more suitable for a library with basic vector and matrix objects
Hi What generator did you select in cmake? You need to select a generator corresponding...
Hi What generator did you select in cmake. You need to select a generator corresponding...
Hi Instructions specific for MSVC++ are found here: http://itpp.sourceforge.net/4.3.1/installation.html#inst_instr_cmake...
Hi We should update the IT++ code to C++11. This task might require a moderate rewrite...
Hi We should update the IT++ code to C++11. This task might require a moderate rewrite...
Hi The ideea is very good especially because we get travis-ci support. On how many...
Please follow the instructions found in install.txt file
LDPC_Generator_Systematic::construct() returns wrong ordering
changes pushed into master
solved bug in LDPC class
Hi IT+ uses cmake compilation system to compile and install IT++. Please follow the...
Hi For matrix concatenation look at Mat class help http://itpp.sourceforge.net/4.3.1/classitpp_1_1Mat.html...
Check cmake documemtation. Also github projects might help https://github.com/sz...
Hi IT++ relies now on cmake for compilation and I'd recomment using cmake when you...
what Linux distribution are you using ? MKL version ?
Hi There are actually 2 types of *.lib: one is the static library the other one is...
Hi The static version of the library is just an archive of the object files of the...
Hi You should try to modify FindBLAS.cmake file, then use the cmake compilation system...
Hi You need to put that library in a folder that us listed in PATH environment variable....
Hi Yes, they look like dependences from external libraries. regards Bogdan
Add cmake example for cross-compiling IT++
Hi It seems that you don't use correctly the 'itpp-config' script. For the correct...
use delete[] for arrays insted of simple delete
using delete[] for arrays (bug ID #248)
Array is accessed out of renge in itpp::lerouxguegenrc
Algorithm computes values not necessary for eva...
Hi My simulations showed that there is an improvement when using more BSs. Here is...
Hi My simulation showed that there is an improvement when using more BSs. Here is...
Hi Indeed, the algorithm needs to be modified after the 2 possible solutions have...
this is an error at link time: you need to link with IT++ library
Hi Could you check that the files have been installed ? From the messages it looks...
Hi You can initialize a matrix from a C-style array: Mat (const Num_T *c_array, int...
I have changed the bug status.
Reed-Solomon Decoder resizes result if non-syst./dec.failure
Hi Thanks Stephan for helping with this. For the workflow just remember to update...
I mean a fully compilable example.
Hi Could you provide a full example in order to debug this issue ? thanks Bogdan
you should check in your project how the GSL library is linked. All errors you have...
Hi Larry Please check the following things: did you provide a root password for the...
Changes have been committed into master. Thanks Etienne for your patch, but please...
LDPC_Code segfault in destructor
Changes have been committed into master. Thanks Etienne for your parch, but please...
solved LDPC_Code segfault in destructor (bug #245)
Hi Jerry NIST Statistical Tests Suite project you found does not need IT++ to be...
Any large number will do, for example: std::numeric_limits<double>::infinity() For...