Measuring zmq latency between two python processes in high resolution in windows 10

I want to measure the latency of zeromq between two python processes in high resolution(in a us/ns unit)(windows 10).

Here's my code.

sub.py

import sysimport zmqimport timeport = "5556"if len(sys.argv) > 1:    port =  sys.argv[1]    int(port)if len(sys.argv) > 2:    port1 =  sys.argv[2]    int(port1)# Socket to talk to servercontext = zmq.Context()socket = context.socket(zmq.SUB)print('connecting to publisher')socket.connect ("tcp://localhost:%s" % port)if len(sys.argv) > 2:    socket.connect ("tcp://localhost:%s" % port1)topicfilter = "10001"socket.setsockopt_string(zmq.SUBSCRIBE, topicfilter)total_value = 0while True:    string = socket.recv()    got_time = time.time() # resolution in windows 10: 15.6ms    topic, msgdata = string.split()    dur = got_time - float(msgdata)    print('it took %.5fms from pub' % (dur*1000))

pub.py

import zmqimport randomimport sysimport timeport = "5556"if len(sys.argv) > 1:    port =  sys.argv[1]    int(port)context = zmq.Context()socket = context.socket(zmq.PUB)socket.bind("tcp://*:%s" % port)topic = 10001while True:    msgdata = time.time() # resolution in windows 10: 15.6ms    socket.send_string("%d %.5f" % (topic, msgdata))    print("topic:%d, msg:%.5f" % (topic, msgdata))    time.sleep(1)

However, since time.time() function's resolution in windows 10 is only 15.6ms(windows default timer period), I was unable to measure the latency below 15.6ms with code.

So I thought about using time.perf_counter() function, which have microsecond resolution, but it was also useless in measuring time difference of two process since time.perf_counter() function's reference point is undefined according to doc.

Is there any other way to measure the time difference between two python processes in high resolution(microsecond or nanosecond unit)?