This project represents a the official TCP Tx Benchmark for the PicoTCP library.

The mbed board is sending one Megabyte to the other side and we measure how fast it does that.

The result:
PicoTCP TX Throughput: 1048576 bytes in < 1 seconds (10.097)Mbits/s

The python script used (which you should rename it to PicoTCP_Official_TCP_Tx_Benchmark.py)

#!/usr/bin/python
import sys, socket
import random, string
import select
from time import time

ECHO_SERVER_ADDRESS = "192.168.100.12"
ECHO_SERVER_PORT = 7
N_PACKETS = 1024
LEN_PACKET = 1024

TOT_BITS = (LEN_PACKET * N_PACKETS * 8) # TX bits 
PACKET = ''.join(random.choice(string.ascii_uppercase+string.digits) for _ in range(LEN_PACKET))
MEGA = 1024 * 1024.
UPDATE_STEP = (N_PACKETS/20) # Make the update step such as one step = 5%
UPDATE_R_STEP = (N_PACKETS/100) #Make the update step such as one step = 1%
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((ECHO_SERVER_ADDRESS, ECHO_SERVER_PORT))
start = time()
rx_total = 0
tx_total = 0
data = ''
i = 0
lp = []
p = select.poll()
while(True): # keeping this in for robustness testing, when the test will run for days
	#Stage II - receving data from the mbed board
	print "Started receiving data...."
	while(len(data) < N_PACKETS * LEN_PACKET):	
		p.register(s, select.POLLIN)
		lp = p.poll(1)
		if(len(lp) > 0):
		    r = s.recv(LEN_PACKET)
		    if (r != ''):
		      data += r
		      if (len(data) / LEN_PACKET) % UPDATE_R_STEP: print "recvd %d/%d (%.2f%%)" % (len(data)/LEN_PACKET, N_PACKETS, float(len(data))/ float(N_PACKETS * LEN_PACKET) * 100.)
		    else:
		      print "Error receiving !"
		      sys.exit(1)

	print "Test was finished!"
	break;	

t = time() - start
s.close()
print "PicoTCP TX Throughput: %d bytes in %d seconds (%.3f)Mbits/s" % (TOT_BITS / 8, t, ((TOT_BITS / t) / MEGA))

This test is based on the following libraries