Add support for compiling on other architectures (DSP-172)

[coderforlife]

Is your feature request related to a problem?

Currently the library can used for testing on Linux with x86_64 due to the code at

esp-dsp/modules/common/include/dsp_common.h

Lines 86 to 88 in 3c12d05

	// Linux Target
	#include <x86intrin.h>
	#define dsp_get_cpu_cycle_count __rdtsc

.

Describe the solution you'd like.

However, this could be easily adjusted to support other architectures by using some of the following:

• If the compiler is clang, it always has __builtin_readcyclecounter() which works on any architecture that supports it (x86, ARM, etc). On x86_64 it is equivalent to __rdtsc().
• Using PMCCNTR on most ARM architectures (but only under certain circumstances).
• Falling back to the vDSO clock_gettime(CLOCK_MONOTONIC) which is slightly slower than __rdtsc() due to conversion to nanoseconds.

A very complete system is outlined at https://github.com/google/benchmark/blob/v1.1.0/src/cycleclock.h but that is likely overkill.

Another solution would be to disable all timing codes if necessary.

Describe alternatives you've considered.

I modified the dsp_common.h header myself to use __builtin_readcyclecounter() in my environment and to not include the intrinsics file (which is the real problem on other architectures).

Additional context.

No response