We have #glibc, #zlib, #gmp, #mpfr and #mpc ... in theory everything needed to build a full-featured native #gcc for C and C++. Oh wow. Now, trying to create *this* port 😎

Anmeldefrist endet am 31.05.2023!
MPIfR: MoonBounce – Eine Reise zum Mond und zurück
Im Wissenschaftsjahr 2023 – „Unser Universum“ schicken Schulkinder Signale zum Mond und zurück zur Erde. Eine Pressemeldung des Max-Planck-Instituts für Radioastronomie, Bonn. (sg) #MPfR #moonbounce
https://www.raumfahrer.net/mpifr-moonbounce-eine-reise-zum-mond-und-zurueck/

GNU MPFR 4.1.1 is released. https://www.mpfr.org/mpfr-4.1.1/ #mpfr

Updated my build instructions for #GMP / #MPFR / #MPC using #Emscripten, prompted by github stalebot commenting on an issue I had a year ago.

Seems 64bit works now (no need for 32bit chroot, and anyway it seems Emscripten no longer supports 32bit afaict?) so that simplifies things.

It got faster too, now the JS in Node runs only 15x slower than native, rather than 90x slower than native.

https://stackoverflow.com/a/43583154

rounded is a Haskell package for correctly-rounded arbitrary-precision floating-point arithmetic, binding to the MPFR C library. I took over the maintainership last year.

The next releases will be twofold, with branches for MPFR 3.1 (rounded-0.x) and MPFR 4.0 (rounded-1.x). Users should depend on the lowest version(s) that have all the features and bug-fixes they need.

Changes since the previous release are a bug-fix for an embarrassing crash in the long double support, and wrappers for more of the MPFR API.

You can check the release candidates at:
https://hackage.haskell.org/package/rounded-0.2/candidate
https://hackage.haskell.org/package/rounded-1.0/candidate

I use rounded in et, an escape time fractal explorer, as well as other projects. The image was made with et.

#haskell #mpfr #rounded #FloatingPoint #ReleaseCandidate #et

The GNU #MPFR 4.0.2 release candidate is available. https://sympa.inria.fr/sympa/arc/mpfr/2019-01/msg00001.html

finally got around to releasing

https://hackage.haskell.org/package/rounded-0.1

rounded: Correctly-rounded arbitrary-precision floating-point arithmetic (a #haskell binding to #MPFR )

originally by ekmett but abandoned due to technical issues, after #ghc integer-gmp library changed internally I could rewrite some of it as a more straightforward library binding and got it working (that was mostly before 2017)

today I tidied up a few loose ends, added long double support, and pushed the publish button.

On a #Perl script with #MPFR computations in small/medium precision via Math::MPFR, I have noticed a speedup of 18% between #Debian 8 (jessie) and Debian/unstable, mostly due to Perl overhead!
Note: I have recompiled GMP, MPFR and Math::MPFR with GCC 4.9 on both machines. On a C program with a MPFR computation in a large precision, there is a 5% speedup, but I don't know why (both machines have the same processors at the same frequency, and AFAIK, both are HP Z800 machines).