zimmermann6
zimmermann6
with Openlibm 0.7.5 I still see the following issues for long double functions: * a somewhat large maximal known error for atanhl (85.4 ulps) * huge error for lgammal (9.08e19...
I am about to release a new update of the accuracy comparison of mathematical functions. It will be with Openlibm 0.7.5, the last release. The previous update (with Openlibm 0.7.4)...
with the attached program I get with OpenLibm 0.7.4: $ gcc -fno-builtin test_pow_openlibm.c /localdisk/zimmerma/openlibm-0.7.4/libopenlibm.a -fopenmp $ OMP_NUM_THREADS=1 ./a.out nthreads=1 $ OMP_NUM_THREADS=2 ./a.out nthreads=2 Error for powl(0x8.779021e7c2f81b2p+14982,-0x8.0021b03e1f821c9p-15097), expected 1, got 0x8.8980e8092da8445p-3...
the attached program gives +Inf for x=-0xd.b6e8f5c28f5c29p+7l, whereas glibc gives 0x0.01dbd551da54538p-16385 which is the correctly rounded result. [test_tgamma_openlibm.c.gz](https://github.com/JuliaMath/openlibm/files/5923918/test_tgamma_openlibm.c.gz)
while compiling fplll (revision 7a7d1c1) on a Power8 machine, with gcc 10.2.1, I get compiler warnings: ``` CXX libfplll_la-gso_interface.lo gso_interface.cpp: In member function ‘ZT fplll::MatGSOInterface::get_max_gram() [with ZT = fplll::Z_NR; FT...
the current C working draft [1, p392] has reserved names for correctly rounded functions (cr_exp, cr_log, cr_sin, ...). We propose to provide such correctly rounded implementations for the three IEEE...
I've found several discrepancies between CRLIBM and GNU MPFR, see attached file. [bugs.txt](https://github.com/taschini/crlibm/files/7846404/bugs.txt)
the current C working draft [1, p392] has reserved names for correctly rounded functions (cr_exp, cr_log, cr_sin, ...). We propose to provide such correctly rounded implementations for the three IEEE...
In OpenLibm 0.8.3, the long double expm1l function returns NaN for x=0x2.c5c85fdf170c604cp+12l instead of 0xf.ffffcfce79e56d5p+16380l.