fix illegal memory access

fix CUDA 9 shuffle warning

add NVIDIA Volta support

fix assert in AMD backend

remove restricted pointer

fix bug announced in #52

If the miner is compiled in debug mode the assert condition results in a compile time error.

add `sm_70` if CUDA 9+ is avalable

use `__shffl_snyc` if CUDA 9+ is avalable

add order_fix for double hash

Make hwloc check capabilities and stop it from leaking mem

Re-add UAC manifest and WIN_UAC cmake option

Affinity fix version C

Make sure all singletons are set to null and make env a global ptr

remove the gpu_thread_num safety

Keep consistency between class and struct

fix cuda 7.5 compile

fix CMake find package output

fix error: `identifier "__builtin_ia32_monitorx" is undefined`

remove early returns

improve AMD auto suggestion

Minor doc fixes

fix #3

- remove QUIET from find package for OpenCL and CUDA
- remove pascal architecture from default CUDA_ARCH for CUDA < 8.0

- use always a multiple of the number of compute units
- add more device information to the config file

Early returns within a kernel can have unexpected behavior, this strongly depends on the opencl
runtime compiler. To avoid errors all early returns are removed.
There is no negative effect(e.g.performance) if the threads stay alive up to the end of the kernel.

One source for the early return is: http://al-key-opencl.blogspot.de/2014/09/be-careful-not-to-mix-early-return-and.html

This pull request also fix a race condition where all fill the shared memory.
Also a bug introduced with #16 is fixed, because of the early return not was possible that
the last block works with an wrong initilized shared memory (result should be a wrong hash if the result target size is valid).

avoid possible illegal memory access

fix travis OSX build