Ruby Multi Threading With Pool Size

A continuation from my previous Sending Personalized SMS article. Lets imagine telco allows you to send as many messages as you can without limits and you have 100,000 messages to send to all your customers.

If you can only send 1 message per second, it will take almost 28 hours to complete. What if you run it multi threaded with a pool size of 10? That will shorten it down to almost 3 hours to complete the same task.

Say you have the name list of the 100,000 customers in “list.txt”.

+6598765432|Person1
.
.
.
.
.
+6587654321|Person100000

Ruby codes below reads “list.txt”, push it into the queue. Sends SMS multi threaded with a pool size of 10.

POOL_SIZE = 10

def send(strPhoneNumber, strMessage)
  `./adb shell service call isms 7 i32 0 s16 "null" s16 "#{strPhoneNumber}" s16 "null" s16 "'#{strMessage}'"`
end

def read_name_list
  q = Queue.new
  rows=(`cat list.txt`).split("\n")
  for z in 0..rows.length-1
    row=rows[z]
    cells=row.split('|')
    h=Hash.new
    h['number']=cells[0]
    h['name']=cells[1]
    q.push(h)
  end
  return q
end

q = read_name_list
workers = (POOL_SIZE).times.map do
  Thread.new do
    begin
      while h = q.pop(true)
        message='Hi %s' % h['name']
        send(h['number'],message)
      end
    rescue ThreadError
    end
  end
end

workers.map(&:join)

The functioning codes makes explanation of multi threading with pool messy. Let me try to simplify it. I have a queue with 101 numbers, pool size of 10. Multi threading will work on the items in the queue but not in sequence. If you run the code below, you will notice that the number is not in order.

def do_some_task(param)
  puts param
end

POOL_SIZE = 10

q = Queue.new
101.times{|i| q.push(i)}

workers = (POOL_SIZE).times.map do
  Thread.new do
    begin
      while n = q.pop(true)
        do_some_task(n)
      end
    rescue ThreadError
    end
  end
end

workers.map(&:join)

< Back to all the stories I had written