Optimized cryptocurrency mining: Intel secures a patent
The Santa Clara company is recognized as having a patent that covers technology capable of mining cryptocurrency with lower energy consumption, and with further room for improvement.
Intel has secured a patent related to a processor that is described as capable of performing high-performance cryptocurrent mining in an extremely efficient manner, among other things with an explicit reference to the SHA-256 algorithm used by the main cryptocurrency at the world by market capitalization, namely Bitcoin. The patent application was made last spring.
Recall that the mining process is, in a nutshell, a reverse hashing operation: it is necessary to determine a number (nonce) that allows the SHA-256 hash generated by a set of data (which represents the block of the blockchain that is trying) to be below a certain threshold. The hardware accelerators used in Bitcoin mining are needed for 32-bit nonce processing, and the current Application-Specific Integrated Circuit (ASIC) process all of this at various stages and with redundancies.
The patent reads: ” ASICs dedicated to bitcoin mining are used to implement multiple SHA-256 engines and can deliver performance of thousands of hashes per second, consuming hundreds of watts .. A realization of what described here uses micro-architectural optimizations between where the selective hardwiring of some parameters in Bitcoin mining computations “.
The hardwiring of these parameters would reduce the number of calculations required, precisely avoiding some redundant operations: such a system would reduce by about 15% the energy needed for chips, which would also be smaller than those used for Bitcoin minerals.
Today, the patent also suggests that a change in the way the 32-bit nonce is compared for validity could further reduce the energy requirements: ” Instead of comparing the final result of hashes to the target value, the mining appraisal could determine whether the ‘hash out has a minimum of 0 initial nuances, ” the patent reads.