Tag: Programming

ZeromMQ REQuest/REPly pattern

More zeromq testing. See also PUB/SUB.

This pattern is for a reply&request situation where one machine asks for data from another machine. The pattern enforces an exchange of messages like this:

REQuester always first sends, then receives.

REPlyer always first receives, then sends.

Requester

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

import zmq
import time
 
# request data from this server:port
server = "192.168.3.38"
port = 6000
 
context = zmq.Context()
socket = context.socket(zmq.REQ)
socket.connect ("tcp://%s:%s" % (server,port) )
 
while True:
    tstamp = time.time()
    msg = "%f" % tstamp # just send a timestamp
    # REQuester: TX first, then RX
    socket.send(msg)        # TX first
    rep = socket.recv() # RX second
    done = time.time()
    oneway = float(rep)-tstamp
    twoway = done - tstamp
    print "REQuest ",msg
    print "REPly   %s one-way=%.0f us two-way= %.0f us" % ( rep, 1e6*oneway,1e6*twoway )
    time.sleep(1)
 
# typical output:
# REQuest  1392043172.864768
# REPly   1392043172.865097 one-way=329 us two-way= 1053 us
# REQuest  1392043173.866911
# REPly   1392043173.867331 one-way=420 us two-way= 1179 us
# REQuest  1392043174.869177
# REPly   1392043174.869572 one-way=395 us two-way= 1144 us

The one-way delay here assumes we have machines with synchronized clocks. This might be true to within 1ms or so if both machines are close to the same NTP server.

The two-way delay should be more accurate, since we subtract two time-stamps that both come from the REQuesters system clock. Much better performance can probably be achieved by switching to C or C++.

Replyer:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

import zmq
import time
 
port = 6000
context = zmq.Context()
socket = context.socket(zmq.REP)
# reply to anyone who asks on this port
socket.bind("tcp://*:%s" % port) 
 
print "waiting.."
while True:
    #  Wait for next request from client
    message = socket.recv() # RX first
    print "Received REQuest: ", message
    tstamp = time.time()
    sendmsg = "%f" % tstamp # send back a time-stamp
    print "REPlying: ",sendmsg
    socket.send(sendmsg) # TX second
 
# typical output:
# Received REQuest:  1392043244.206799
# REPlying:  1392043244.207214
# Received REQuest:  1392043245.209014
# REPlying:  1392043245.209458
# Received REQuest:  1392043246.211170
# REPlying:  1392043246.211567

Matrix determinant with Boost::uBLAS

Boost uBLAS provides BLAS functionality, but doesn't have a function for computing the determinant of a matrix. Googling for this turns up a few code snippets, but it's best to document this completely here now since I got it to work, and it will be useful for opencamlib sooner or later.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

#include <boost /numeric/ublas/matrix.hpp>
#include </boost><boost /numeric/ublas/io.hpp>
#include </boost><boost /numeric/ublas/lu.hpp>
 
namespace bnu = boost::numeric::ublas;
 
int determinant_sign(const bnu::permutation_matrix<std ::size_t>& pm)
{
    int pm_sign=1;
    std::size_t size = pm.size();
    for (std::size_t i = 0; i < size; ++i)
        if (i != pm(i))
            pm_sign *= -1.0; // swap_rows would swap a pair of rows here, so we change sign
    return pm_sign;
}
 
double determinant( bnu::matrix<double>& m ) {
    bnu::permutation_matrix</std><std ::size_t> pm(m.size1());
    double det = 1.0;
    if( bnu::lu_factorize(m,pm) ) {
        det = 0.0;
    } else {
        for(int i = 0; i < m.size1(); i++)
            det *= m(i,i); // multiply by elements on diagonal
        det = det * determinant_sign( pm );
    }
    return det;
}
 
int main () {
    bnu::matrix<double> m(3, 3);
    for (unsigned i = 0; i < m.size1() ; ++i) {
        for (unsigned j = 0; j < m.size2() ; ++j) {
            m (i, j) = 3 * i + sqrt(j+1); // fill matrix
            m(i,j) = m(i,j)*m(i,j);       // with some numbers
        }
    }
    std::cout << "before det() call m= " << m << std::endl;
    double det = determinant(m);
    std::cout << "after det() call  m= " << m << std::endl; // m has changed afted determinant() call!
    std::cout << "determinant=" << det << std::endl;
}

I'm trying the WP-syntax plugin here for the first time. The include statements are garbled, but otherwise it seems to work.

download source: utst1.cpp

This compiles on Lucid Lynx with the following CMakeLists.txt:

cmake_minimum_required(VERSION 2.6)
PROJECT(utst1)
find_package( Boost )
if(Boost_FOUND)
    include_directories(${Boost_INCLUDE_DIRS})
endif()
ADD_EXECUTABLE(utst1 utst1.cpp)
target_link_libraries(utst1 ${Boost_LIBRARIES})

Composite cutters for ocl

People who, unlike me, actually know something about programming often talk about design patterns. One common idea is to compose objects out of other objects.

I was able to add four new APT-tool like cutter classes to ocl with about 5-lines of code for each cutter (sans the bugfixing, taking much longer, that also took place simultaneously 🙂 ).

These show CL-points resulting from the vertexDrop() function, which results in a shape that looks like the cutter, but inverted.

Other combinations of the basic shapes (cylinder, sphere, torus, cone) are fairly easy to add now also if someone actually needs them.