The Tesla V100 with its new dedicated tensor cores was mainly designed for deep learning. However, since memory has always been the greatest bottleneck in GPU computing, the major performance boost of almost all applications is due to the new stacked memory DRAM used for the Tesla P100, V100 and Titan V models which provide transfer rates of up to 900 GB/s (compared to <400 GB/s for GDDR5X memory modules).

If you take a close look at the numbers you can see that the hashrate speedup roughly resembles the increased memory transfer rate.