Remove legacy code for supporting an old Arm Optimised Routines
deprecated feature for throwing SIMD Exceptions.

Reviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>

Load the polynomial evaluation coefficients into 2 vectors and use lanewise MLAs.
Also use intrinsics instead of native operations.
expf: 3% improvement in throughput microbenchmark on Neoverse V1, exp2f: 5%,
exp10f: 13%, coshf: 14%.

Reviewed-by: Wilco Dijkstra  <Wilco.Dijkstra@arm.com>

Compilers may emit calls to 'half-width' routines (two-lane
single-precision variants). These have been added in the form of
wrappers around the full-width versions, where the low half of the
vector is simply duplicated. This will perform poorly when one lane
triggers the special-case handler, as there will be a redundant call
to the scalar version, however this is expected to be rare at Ofast.

Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>

Some routines reuse table from v_exp_data.c