sleef
sleef copied to clipboard
shibatch
SIMD Library for Evaluating Elementary Functions, vectorized libm and DFT
Add a new workflow to detect unexpected backward-incompatible changes at precommit, by comparing current version against a reference version, both generated on Linux with gcc. It uses abidiff to compare...
Libm and quad tester rely on Unix specific sys/wait.h for pipes. This limits testing capacity on Windows, e.g. with msys2/mingw, and other non-unix based OS-es. Fixing this would allow to...
# What is the purpose of this pull request? Check if existent micro benchmark program is working. The micro benchmark system in place would compile and run hitting no errors,...
**Describe the bug** The `Sleef_atanhd1_u10purec` function is returning values with an error of slightly more than 1ULP for arguments around 10^-17. I've also experienced similar issues with other atanh functions,...
**Where?** https://sleef.org/2-references/libm/#sleef_acosh_u10 **What?** This is a very minor point, but in the description for `acosh` it states ranges of `[-1.34e+154, 1.34e+154]` for double precision and `[-1.84e+19, 1.84e+19]` for single precision....
As described in https://github.com/shibatch/sleef/issues/544, I am now considering to incorporate [tlfloat](https://github.com/shibatch/tlfloat) as a submodule of sleef. The primary reason for incorporating tlfloat is to provide an alternative tester to tester1,...
On S/390 I ran into problem with the way the VXE2 hardware feature is currently being detected in multi-threaded code. It uses a SIGILL signal handler and sigjmp/longjmp in order...
This implements a suggestion from @shibatch for #561. It uses a pthread mutex to prevent multiple threads from accessing the sigjmp variable in the feature detection code simultaneously. Remaining todos:...
I need to use the quadruple-precision floating point number in CUDA. The Quad-precision math library reference for CUDA [shows](https://sleef.org/quadcuda.xhtml) that I should use _**sleefquadinline_cuda.h**_ and the type _**Sleef_quadx1**_. I have:...
**Describe the bug** This issue is about a bug in gcc-13 that was already addressed in https://github.com/shibatch/sleef/pull/592. When compiling with gcc-13, the sign of signed zero sometimes seems to be...