Is is possible that the number of results are greater than the result output array,
in this case invalid memory can be access within the device and on the host side.

fix invalid kernel call parameter

Port the crash fix from xmr-stak-cpu

check that the number of global threads is a multiple of the workgroup size

Fix large pages on Windows and make them more user-friendly

Increase the scratchpad size to make sure we don't malform JSON

fix amd auto suggestion

fix broken nicehash support

Major cleanup and speedup of soft aes

The number of threads within a kernel must be a multiple of the worksize.
If not it can crash on some systems.

Platform index zero was always used to search for the device detection.

Fix: use correct detected AMD platform index

Merge branch 'topic-large-pages-win' of https://github.com/fireice-uk/xmr-stak into topic-large-pages-win

Topic cleanup

- add nicehash flag back to `miner_work`

- Add a proper UAC manifest requesting elevation when we start
- Fix self_test() - this is integral part of initialisation

readd nicehash pool support

use nicehash start nonce in all backends if nicehash is enabled

fix donation section

refactor xmr-stak

- add CIvariables to debug cuda build

- fix windows CI
- add CUDA to linux CI

- add more documentation
- split documentation into different files

- string literal can only be 16kbyte