I am at 1080ti maximum reached 380 Mkey / s. This is probably normal relative to 1070 and AMD. I wonder why 2080ti is more powerful in ~ 3 times than 1080ti, it should be about ~ 30% (500 Mkey / s). Due to what such a gap, have thoughts?
RTX 2080ti has 4,352 data flow processors, GTX 1080ti has 3,584 data flow processors. So, it is reasonable that GTX 1080ti should have at least 800-900 Mkey/sec. If we have a look through "market price" terms, so the minimum should be 500-600 MKey/sec. Can you try to change and test different parameters b, t, p for block, threads and points? The number b (blocks) try to put 2-3 times more than the number of computing items of your card.
But anyway, thank you for your answer. The previous one I had was just 150 MKey/sec with GTX 1080ti whch seems totally insufficient and wrong.