fix high cpu usage on mtp algo

fix high cpu usage for mtp (both pool and solo).
cpu usage stays around 0-0.1% as it was before the changes in gpu driver of last year.

support for 30x with ccminer_cuda11.exe

stability improvement

corrects a memory leak which may cause ccminer after a long while if the ressources (ram/virtual memory) are limited.

Procedure for solo mining.

wallet side
edit or create the file zcoin.conf and add these informations to allow mining (this is an example):
server=1
rpcuser=djm34
rpcpassword=password
rpcport=8382
rpcallowip=127.0.0.1 (if the IP address of the miner is the same as the one of the wallet), otherwise it should be the address of the miner)

in the miner command line:
ccminer -a mtp -o http://127.0.0.1:8382/ -u djm34 -p password --coinbase-addr zcoin_address --no-getwork --no-stratum

In the case of the wallet is on a different machine http:// should point to the IP of the machine where the wallet is located and in the wallet config, rpcallowip to the address of machine where the miner is located

The address given in --coinbase-addr, does not need to be an address of the wallet to which the miner is connected any zcoin address should work.

fix solo mining

This release fix a bug in solo mining which was using the extranonce increment which is only for pool mining

improved stability

• Correct an issue which may cause a crash on start or when the stratum_thread is restarted (ie: change of pool, restart of pool connection etc...)
  This problem was noticed on linux, but hasn't been observed on windows so far

fix solo mining

• correct issues with solo mining which now works properly

update to solo mode

solo mode now use same initialization and parameter as the pool mode.

better handling of eventual packet losses

• better handling of eventual packet losses when the connection to the pool is less than optimal

stability improvement

rearrange how the cards are assigned to thread. this should correct the issue described here #48

stability improvement

correct more crash issues...
These crash were occurring when the pool stop sending data or is unable to send data.

When this happen, the stratum thread is restarted.
In case of vardiff, the last difficulty is used when the thread is restarted.

solved crash problem

• better handling of miner/pool connection:
  In case a pool connection problem, the miner restart the stratum_thread on the current pool.
  if a failover has been set, it will restart on the failover

• solve various crash problem due to pool sending empty data

• sliglty improved speed on 20x cards