Redis [QUESTION] Is it feasible to split all commands into two parts to improve performance?

Is it feasible to split all commands into two parts to improve performance? The first part, which includes parameter validation and content checks (non-memory operations), could be processed in the I/O thread, while the second part involving concurrency-sensitive operations would be handled in the main thread.

example(And maybe can split setGenericCommand?):

void setCommand(client *c) {
    robj *expire = NULL;
    int unit = UNIT_SECONDS;
    int flags = OBJ_NO_FLAGS;
========================= proc in iothread ===============================
    if (parseExtendedStringArgumentsOrReply(c,&flags,&unit,&expire,COMMAND_SET) != C_OK) {
        return;
    }

    c->argv[2] = tryObjectEncoding(c->argv[2]);
========================= proc in iothread ===============================
    setGenericCommand(c,flags,c->argv[1],&(c->argv[2]),expire,unit,NULL,NULL);
}

Comment From: wclmxxs

@sundb @ShooterIT What do you think?

Comment From: sundb

@wclmxxs Theoretically, it is feasible, but how much benefit will it actually bring? Argument parsing is generally not a bottleneck. We can make a POC to verify it.

Comment From: wclmxxs

@sundb I want to make sure that I can use ./memtier_benchmark --data-size 521 --ratio 1:1 --key-pattern R:R --key-minimum=1 --key-maximum 3000000 --test-time 60 -c 50 -t 13 --hide-histogram -x 3 for performance testing.

Comment From: sundb

@wclmxxs, sure. Additionally, you can use the --command option to specify a specific command. I think it is more suitable for this case.