We did some improvements on our EQM miner.
Only adjusted for Pascal cards so far. Report your Maxwell speeds.
GTX 1070: ~125 Sol/s @ 110 W
GTX 1060 6GB: ~85 Sol/s @ 80 W
GTX 1050 Ti: ~53 Sol/s @ 45 W
GTX 1050: ~48 Sol/s @ 40 W
GTX 980 Ti: ~125 Sol/s
GTX 980 OC: ~115 Sol/s
GTX 980: ~100 Sol/s
GTX 970: ~85 Sol/s
GTX 950: ~50 Sol/s
GTX 750 Ti: ~36 Sol/s
GTX 750 1GB OC: ~34 Sol/s
This miner is locked to NiceHash and does not have any dev fee. If you do not agree with that, do not use it.
Windows x64 (Maxwell and Pascal cards, CPU AVX/AVX2): EQM v1.0.1ba
CPU mining: this miner has support for LOCKED PAGES; it can give few extra % of performance. How to enable: https://msdn.microsoft.com/en-us/library/ms190730.aspx (you may have to restart OS after changing). CPU mining supports "big servers" with more than 64 cores and there is no need to manually fiddle with affinities and NUMA nodes. Just set -p (number of CPUs) and -t (number of CORES per CPU) and miner automatically deals with it.
Examples of usage
Query CUDA devices:
Benchmark device 0 for 10 seconds:
eqm -t 0 -b 10 -cd 0
Mine with device 0 on EU server and set worker name:
eqm -u YOURBTCADDRESS -w worker1 -cd 0
Mine with devices 0 and 1 on USA server:
eqm -l usa -u YOURBTCADDRESS -cd 0 1
Launch with multiple threads/instances per CUDA device:
eqm -t 0 -b 10 -cd 0 0