Kim Walisch

Sorry, but this is currently not supported. It should not be too difficult to modify calculator.hpp to support double and float but you need to do it yourself. I could...

I have benchmarked ```WITH_DIV32 ``` on an Intel i5-8400 CPU & Ubuntu x64: * clang++-14 & ```WITH_DIV32=OFF```: ``` $ ./primecount 1e15 -l -s === pi_legendre(x) === pi(x) = phi(x, a)...

And here is the same benchmark using GCC11 on my latest Intel i5-12600 Alder Lake CPU from 2021: * g++-11 & WITH_DIV32=ON: ``` $ ./primecount 1e16 -s -l === pi_legendre(x)...

> Should this be added to docs? Personally I feel like the documentation should not contain too many expert topics. My documentation is currently written so that beginners and intermediate...

Sorry for not answering earlier. Based on my experience I think your MINGW installation is corrupted, in primesieve's CI MinGW64 is tested and works flawlessly. Also have a look at...

Both GCC & Clang now optimize integer division by default, hence I have now disable ```WITH_DIV32```by default in ```CMakeLists.txt```: https://github.com/kimwalisch/primecount/commit/00dcb71b5e7008da5c5e9f48a47a26317d0b6c4c

Thanks to the LLVM devs for causing useless work in downstream projects! Apparently this warning only applies to C99 through C17, from C23 on having a function declaration without a...

Yes, I know about this issue. If you are on Linux and you use the GCC compiler, then you can build primecount using 128-bit floating point numbers which increases the...

> Maybe adding a comment there would make sense. I have done that: https://github.com/kimwalisch/primecount/commit/fdecedce6dfc58ecc353df8f015460c65ec8892b > What stops cmake from choosing WITH_FLOAT128=ON where possible? I did some research and found a...

I did not even know CMake supported unity builds until now. I have now tried CMake unity builds of these older primesieve versions: 11.1, 11.0, 8.0, 7.9 and 7.0 (which...