Optimize locks to avoid full iteration of lock map#2701

gztensor wants to merge 5 commits into

devnet-readyopentensor/subtensor:devnet-readyfrom

feat/reverse-lock-mapopentensor/subtensor:feat/reverse-lock-mapCopy head branch name to clipboard