xtensor
xtensor copied to clipboard
C++ tensors with broadcasting and lazy computing
Windows 10 MSVC 2019 Enterprise C++20 Fails to compile with both XSIMD and TBB enabled. Fails with the following `syntax error: 'template'` On line 1252 of xassign.hpp
just to not forget this work :) One question: should this better live at the xsimd level?
Hi, I am still getting used to the library, but was able to isolate an unexpected performance hit. I want to update just a subregion of a pre-allocated 1D tensor....
I have added a couple routines with doxygen but no reference in readthedocs.io, there are likely many functions missing from the package.
It would be useful to have an equivalent function in xtensor to ndindex. https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndindex.html This allows to iterate over all the nd indices of a shape regardless of the number...
xsimd 10.0.0 exposed an error from xtensor (0.24.7) on arm64 architecture (neon64). The error was reported at https://github.com/xtensor-stack/xsimd/issues/945 : ``` 127s [ 26%] Building CXX object CMakeFiles/test_xtensor_lib.dir/test_xoptional_assembly.cpp.o 127s /usr/bin/g++ -DXSIMD_ENABLE_XTL_COMPLEX...
The following simple test code does not compile when using -DXTENSOR_USE_XSIMD: ``` #include void xtensor_or(xt::xtensor_fixed& b1, const xt::xtensor_fixed& b2) { b1 = b1 | b2; } ``` The compiler complains...
The current documentation is "auto data() -> ... Return the pointer to the underlying buffer." For people using this library for the first time it may be easy to misunderstand...
#2435 (PR #2436) fixed some instances of EXPECT_EQ being used for floating point comparison instead of EXPECT_DOUBLE_EQ. gcc-13 appears to have uncovered a few more, causing test failure on i386...
[Question] Is there an easy path to point the XSIMD allocator at std::pmr or a custom memory pool?
I want to be more flexible in my memory usage in my program's hot path. To do this I need to point xtensor allocations at a memory pool ideally also...