XRTRunner Utility Class & Programming Examples Cleanup

[hunhoffe]

This PR:

• Introduces the XRTRunner class, which is used to run programs on the npu while using common numpy operations to check the input/output
• Remove spaghetti python path logic in examples
• Allows the data_transfer_transpose to run for both float and int 32-bit datatypes
• Makes it easier to change the sizes of things in the matrix_scalar_add example
• Percolate changes from Fixup data layout in programming examples #672 to other programming examples as appropriate
• Remove relative paths in makefiles
• Previously, the matrix_scalar_add/multi_launch_channel example would not compile; but then a recent PR (Fixup data layout in programming examples #672) fixed that issue. However, the code as-is was in an intermediate state (testing something simple before full implementation) so it didn't truly fix the issue. I've updated the code so the implementation is correct; we now get a failure with the complete code. I've marked the test as a failure and will re-open the multi-launch issue (Multiple launches, herd with single core #627)
• Adds/updates READMEs

The XRTRunner class does truly reduce the number of lines of code; however, as I improved the ability to configure examples (and add documentation, etc.) these things added back a lot of lines of code to this PR so the improvement isn't immediately obvious in the git LoC counts.

[fifield]

Looks good, thanks!