aarch64 support

[christophe0606]

Some build issues reported for cortex-a76 / cortex-a55:

qsub intrinsics not supported. But the CMSIS Core_A/Include has not compilation flag to disable the definition of intrinsics like qsub.

In arm_max_f32 and arm_min_f32:

/home/sandal/CMSIS_5_BK/CMSIS/DSP/Source/StatisticsFunctions/arm_max_f32.c:168:24: warning: conversion from ‘long unsigned int’ to ‘uint32_t’ {aka ‘unsigned int’} changes value from ‘18446744073709551615’ to
4294967295’ [-Woverflow]
  168 |   maxIdx = vdupq_n_u32(ULONG_MAX);

ULONG_MAX is 32 bits so I don't understand why the value is 64 bits according to the compiler.

[ua1arn]

ULONG_MAX in both places need to replace with UINT32_MAX

[christophe0606]

@ua1arn Thanks. Already done in a recent update :-)