* Add basic rotation benchmarks

* Rename benchmark name

* Add Resize benchmarks

* Add supersampling exceptions

* Adjust bench angle for non special case

* Add integer math helpers

* Add cache blocking to rotate

* Add large image rotation bench

* Change cache size

* Optimize repeated calculation

* Roll back to no-cache version

* Fix tests

* Add doc and tests

* Cleanup

* Add convolve benchmarks

* Partially added Pad and Extend

* Implement padding methods

* Reduce allocations

* Improve performance of convolve function

* Add more image data

* Refactor convolve operation and fix output padding

* Re-enable parallel convolution

* Refactor convolve and execute functions

* Simplify implementation

* Add tests

* Add keep alpha

* Improve doc

* Fix alpha pos

* Improve tests