The following macros are removed:
* EIGEN_DECLARE_CONST_Packet8f
* EIGEN_DECLARE_CONST_Packet4d
* EIGEN_DECLARE_CONST_Packet8f_FROM_INT
* EIGEN_DECLARE_CONST_Packet8i
The following preprocessor macros are added:
- EIGEN_COMP_CPE and EIGEN_COMP_CLANGCPE version number of the CRAY compiler if
Eigen is compiled with the Cray C++ compiler, 0 otherwise.
- EIGEN_COMP_FCC and EIGEN_COMP_CLANGFCC version number of the FCC compiler if
Eigen is compiled with the Fujitsu C++ compiler, 0 otherwise
- EIGEN_COMP_CLANGICC version number of the ICX compiler if Eigen is compiled
with the Intel oneAPI C++ compiler, 0 otherwise
All three compilers (Cray, Fujitsu, Intel) offer a traditional and a Clang-based
frontend. This is distinguished by the CLANG prefix.
Some some header guards were repeated between the `AltiVec` package and the
`ZVector` packages. This could cause a problem if (for whatever reason) someone
attempts to include headers for both architectures.
1. Speed up exp(x) by reducing the polynomial approximant from degree 7 to
degree 6. With exactly representable coefficients computed by the Sollya tool,
this still gives a maximum relative error of 1 ulp, i.e. faithfully rounded, for
arguments where exp(x) is a normalized float. This change results in a speedup
of about 4% for AVX2.
2. Extend the range where exp(x) returns a non-zero result to from ~[-88;88] to
~[-104;88] i.e. return denormalized values for large negative arguments instead
of zero. Compared to exp<double>(x) the denormalized results gradually decrease
in accuracy down to 0.033 relative error for arguments around x = -104 where
exp(x) is ~std::numeric<float>::denorm_min(). This is expected and acceptable.
Makes e. g. matrix multiplication 2x faster:
name old cpu/op new cpu/op delta
BM_convers 181ms ± 1% 62ms ± 9% -65.82% (p=0.016 n=4+5)
Tested on all possible input values (not adding tests, since they
take a long time).
