Dr. Thomas Orgis

Results 55 comments of Dr. Thomas Orgis

> Also, DSYEVR wasn't changed since [77a7afb](https://github.com/Reference-LAPACK/lapack/commit/77a7afb48d8633d9865a6f7473d542dc81d9bc97), so it is remained not being thread-safe. ``` $ FFLAGS=-Os cmake ../ […] -- Performing Test _recursiveFlag -- Performing Test _recursiveFlag - Failed...

> > Which happens to be named blas.pc in your case, not openblas.pc? I don't see how you can rely on that. > > You cannot rely on it. Exactly...

I have similar trouble trying to package things with pkgsrc. I would like to have a complete install of the reference, with cblas and lapacke. For differing implementations being installed...

PS: With Intel MKL, there is a central switch `-DMKL_ILP64` to be set. I imagine setting up trivial `include/intel-mkl64/cblas.h` with ``` #ifndef MKL_ILP64 #define MKL_ILP64 #endif #include ``` to fit...

> Yes. I agree with that, and prefer the solution that does not replicate the entire header. I think it is cleaner. This is ambiguous to me. Which is the...

A beautiful aspect of shipping/installing the 32 bit cblas.h with this modification to the usual location is that the original mechanics still work. Only the 64 bit variant will enforce...

Oh, come on … the CBLAS/cmake/cblas-config-install.cmake.in seems to forget -DCMAKE_INSTALL_INCLUDEDIR, doesn't it? ``` # Report lapacke header search locations. set(CBLAS_INCLUDE_DIRS ${_CBLAS_PREFIX}/include) ``` (The comment is sugar on top.) I have...

I have to give up now … I managed to move cblas.h to cblas.h.in as indicated above, and added ``` configure_file(${CMAKE_CURRENT_SOURCE_DIR}/cblas.h.in cblas.h @ONLY) configure_file(${CMAKE_CURRENT_SOURCE_DIR}/cblas_f77.h.in cblas_f77.h @ONLY) ``` to CBLAS/include/CMakeLists.txt, also...

@weslleyspereira So you at first liked this idea: ``` #if defined(WeirdNEC) #define WeirdNEC #endif #include "../cblas.h" ``` with /prefix/include/cblas.h and /prefix/include/netlib64/cblas.h, the latter locating the former? But you do agree...

Anything? I must admit that I don't see much chance for a different solution in practice, as this is the example set by openblas, the main implementation we use. I...