Jérôme Richard

Hello, Please read this post https://stackoverflow.com/questions/70311592/numba-np-convolve-really-slow/70312564 . It shows that there are some **issues with the current internal dot product implementation**. Typically, the current dot product BLAS wrapper (combined the...

### Using GCC The number of threads actually created is always one for GCC-8.3 when `OMP_PLACES` is *not* set to `"cores(...)"`. This is not the case with Clang-9.0 (the number...

Hello, I did the same experiment to learn/test ISCP and found a workaround using two compilation flags. The code is bounded by the latency of a floating-point addition (on nearly...

The problem is located in `pop_char` of `fson.f90`: the instruction `str = str(2:)` copy the string without the first character and add a space character at the end (the size...