underlayer 2.01

underlayer is a symmetric decentralized parallel computing and load balancing framework.

  Add it to your Download Basket!

 Add it to your Watch List!


Rate it!
send us
an update
GPL (GNU General Public License) 
3.8/5 14
Peter Alvaro
ROOT \ System \ Clustering and Distributed Networks
underlayer is a symmetric decentralized parallel computing and load balancing framework, providing c, perl and java APIs for distributed job management.


>make linux-shared
>make install

to get the perl API going:

>cd Underlayer/
>perl Makefile.PL
>make install

to get the Java API going:
>cd java

edit make_java.sh to point to your jdk.
>sh make_java.sh

to start the underlayer daemon:

to query the underlayer:
>ul_request [service_id]

to register an already running service at port :

to start a new abstract listening service:

c api:
see the underlayer man page

c parallel programming:
see example1, pm.c

perl api:
see the Underlayer man page

perl parallel programming:
see example2, master_example.plx

perl parallel programming in detail:

create a working directory:
>mkdir work
>cd work

start a listener for service_id 19 at port 1919.
ideally, do this on as many servers running an abstract machine
at service_id 19 as possible.
>abstract 1919

change to the examples directory
>cd ../examples

given a large phrase list and a small phrase list, at 50 threads...
>perl master_example.plx largelist.txt smalllist.txt 50 > results.txt

note that if you do this on a single server, it will never get to anywhere near 50 threads. it blocks as soon as the processor is maxxed out. you can check the load state by running

if you built the java library and program, you can use the GUI monitor:

>java ulDisplay

add 20 servers and try again! you'll be pleased with the speed. you can set the serial speed by running only the slave:

>perl slave_example.plx largelist smalllist.txt > results_serial.txt

Last updated on January 10th, 2008

#clustering program #parallel computing framework #java API #c #perl #java #api

Add your review!