This will allow later more gpu candidates.
Note: This is an unfinished work, we keep the previous behavior for now
To finish this, all algos solutions should be migrated and submitted nonces attributes stored.
Its required to handle the different share diff per nonce and fix the possible solved count error (if 1/2 nonces is solved).
fix some algo weird hashrates (like blake)
and reset device between algos, for better accuracy
but this reset doesnt seems enough to bench all algos correctly...
to test on linux, could be a driver issue...
heavy: fix first alloc and indent with tabs...
reduce "false" warnings, and ignore unrelated/small ones <= 1 MB
On windows the gpu memory can be allocated by other processes
+ some cleanup in algos... (free/gpulog)
remains my original lyra2 implementation to fix... (cuda_lyra2.cu)
I guess some kind of memory overflow force the driver to allocate
memory... but was unable to free it without device reset.
Should give same or better than SP and klaus versions
Keep old code for older devices and skein2 compat
Linux perf: 750Ti 78MH/s and GTX 970 260MH/s
Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>
Tanguy Pruvot