Is there any way to delay the application of overclock on the core during dag creation. I know its possible on the memory with -mcdag but Im having issues on the coreclock during dag creation process. I know increasing memory will increase the hashrate but in my case, increasing the coreclock increases the hashrate as well. Weirdly it only happens on my gtx 1070.
Not that weird - the TLB cache is part of the core, and its speed is the limiting factor of the hashrate of GTX1070 with the current DAG sizes. We could add it but not in the next release.
Is there any way to fix this TLB cache problem? I am one of probably many 1070, and 1080Ti users that would be grateful if there would be some kind of fix for this problem in the upcoming releases. We are losing 4-10 mh/s per card now as oppose to 2017.
Tnx.