# Maximal bit size of multipliers using with primorials 13, 14 and 15;
# Default values are 26, 33, 36
# Less values increases prime probability, but sieve generate less candidates
multiplierLimits = ["24", "31", "35"];
Also try use this multiplier limits.
And if possible, use Ubuntu instead of Windows now (10.2.2 version works on windows with 100% cpu load), CUDA version for Windows will be ready in 2-3 days.
Thanks for getting back so quickly! Yes, those are the settings I'm using. The G4400 CPU's are on a Ubuntu dist. My 1080 Ti's get around 0.45-0.5 and my RX 570's get around 0.3.
Will the CUDA version be faster for Nvidia cards on Windows compared to Linux? I'm in the process of getting rid of Windows if possible.