fix avx detection

fix cn_gpu rounding mode selection

If cryptonight_gpu is used in the dev pool and a cryptonight_v8 derived
algorithm than cryptonight_gpu will create wrong results if v8 was
running as non ASM version.

CPU: fix cryptonight_r slowdown

If the miner changed the pool (user pool or dev pool) and the network
block of monero has not changed the miner switch to the non asm version.

Add check if the hash function has changed. If the hash function not
point to the asm version the ASM function is recreated.

update currency pow table

remove wrong documentation

The OS is able to disable avx. In that case our avx detection is wrong.
See xmrig issue: https://github.com/xmrig/xmrig/issues/951

Import fix: https://github.com/xmrig/xmrig/commit/6052da3c4399514c95245b2d342f04baa20958db#diff-92fa29935c136762d35981769ac79f52

thx @xmrig for the notification.

fix opencl and cuda cryptonight_r caching

use double hash for cryptonight_r

OpenCL: fix windows vega and vii

fix masari

Remove fork description for monero.

Remove a warning about unsupported CUDA from an older release.

The original implementation of the cache release and create always new
kernel, this can lead into performance issues and my crashes.

add support for cnR double thread assembler

Since masari increased the block size the miner crashed each time it
gets connected with a masari pool.

This PR extent the possible size of a block to 128 byte and updated the
kernel.

Fix again issues with vega and vii with the latest windows driver.

increase version to 2.10.0

dev pool change

Edited comments in configs, set default verbose to 4

activate nicehash

Support of CryptoNight v8 ReverseWaltz

CN_GPU: Fix old compilers and old kernels (Clang 3.5 tested, without CONFIG_HUGETLBFS)

rebased version of #2261

Added support of CryptoNight v8 Reverse Waltz (named cryptonight_v8_reversewaltz here) - equal to CryptoNight v8 but with 3/4 iterations of CryptoNight v8 and with reversed shuffle operation

We plan to use CryptoNight v8 Reverse Waltz as new PoW algorithm for Graft (graft-project/GraftNetwork#234).

assembly patching

fix wrong logic operator usage

- fix linker issues with CUDA8
- fix device selection

- add cryptonight_r assembly
- fix cryptobight_v8 assembly

fix #2276

OpenCl: fix multi gpu support

Since #2268 the multi gpu support is broken. It is not possible to have more than one opencl context in a process.
Use one OpenCl context for the process instead per GPU.

update version to 2.9.0

cryptonight_r

merge back master -> dev

log level 4 as default

- add nvrtc to windows libraries
- add path to cuda.so for linux