OpenCL: optimize cn_gpu

fix groestl, skein and blake

Optimize cn_gpu

based on the suggestion from @xmrig https://github.com/xmrig/xmrig-amd/commit/db4e169f3a78f273abf89ea8cf5bba7eccf1490b

remove cn_turtle as native POW 

refactor POW definition

cn_gpu: disable auto suggestion for CPU

cryptonight_turtle is only cryptonight_v8 with a different scratchpad,
iteration and mask value.
We are using now the new machanism to describe such derived POWs.

A POW is now defined by a function `f` and three degrees of freedom `f(iteration, scratchpad, mask)`.
`f` is the base algorithm like `cryptonight, cryptonight_gpu`
An easy to pars snytax to write the full POW definition down is: `cryptonight_gpu:0x0000c000:0x00200000:0x001fffc0`

This change make it very easy to integrate the new trend of variate the
number of iteations or the scratchpad size without modifying the full
code.

To avoid that a miner is waisting power with there CPU to mine cn_gpu the
auto suggestion will comment out the config.

OpenCl: use user defined unroll in cn_gpu

OpenCL: fix invalid work group size

OpenCL: fix groestl

@xmrig provided the information that the driver 19.2.1 for vega also
create invalid results if pragma unroll is used for the groestl algo.

- use the user defined unroll
- auto suggestion:
  - only tune for cn_gpu if this is the main user
currency (after a fork)
  - set unroll to 1 for cn_gpu

OpenCl kernel using a larger work group size than configured by the
user to increase the occupancy. Depending on the algorithm and device
the size is limited.

This PR fixes that the user was able to select a invalid work group size.

OpencL: fix cn_gpu

fix missing algorithm name

CUDA. fix static in global kernel

If comp_mode is used the code will not compile.

- fix compile issue
- fix wrong conditions to handle `comp_mode`

During the implementation of `cryptonight_gpu` we missed to add the name
of the algorithm to jpsock.

- add function to translate an algorithm into a name.

Remove `static constexpr` within the global kernel. This is not
supported by all CUDA versions.

update version to 2.8.0

update documentation

coin cleanup

- remove driver warning
- add link to supported card y the ROCm driver

OpenCL: fix Blake hashing

- rename `Intense` to `Lethean`
- remove `masari` and `stellite`, both hase forked to a currently
unsupported algorithm

Windows driver creates wrong code if unroll is used.

OpenCL: fix work size message

CUDA: optimze cn_gpu auto suggestion

OpenCL: use algorithm names instead of number

add ryo algorithm fork

cuda: optimize cn-gpu

Fix message with the maximal allowed worksize if cryptonight_gpu is
used.

optimize the algorithm for cryptonight_gpu autosuggestion

switch with block 6 to cryptonigh_gpu
(https://github.com/ryo-currency/ryo-currency/blob/7bc5dc40864e518753807d04c17d617f728da156/src/cryptonote_config.h#L185)

psychocrypt committed 9 minutes ago
 - use precomuted indicies within the loop
 - `cn_explode_gpu` use all threads to load the state