Add workaround for another macro name collision with Windows 10 SDK winnt.h

Fix missing EXTERNAL declarations in LAPACK TESTING (LAPACK PR 580)

Power10: Fix for SBGEMM

While testing bfloat16 sbgemm kernel, there are some failures for odd value inputs due to updating result for
additional bytes.

Further shortcuts for (small) cases that do not need buffer allocation

Add shortcut for small-size S/D GEMV_N with increments of one

arm64: add the missing d9 register to the clobber list

Refs. numpy/numpy#18422

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>

Remove spurious casts from PPC parameters and fix compilation for older targets

Optimized sgemv_t for small N based on AVX512

POWER10: Fixes for sbgemm kernel

While testing bfloat16 sbgemm kernel, there are some failures
for odd value inputs due to array access beyond the boundary.

riscv64 fixes

The implementation in `riscv64/dot.c` fails the `test_dsdot` test, and
the generic kernel seems to have better precision. Tested on SiFive
FU740 (HiFive Unmatched) and QEMU.

Also see #1469.

Same as #1990.

revert "try to work around gcc update problems" in Homebrew workflow

...as homebrew has dropped at least gcc8 now