Skip to main content

Posts

Showing posts from December, 2003

Online once more

On this last day of this year, I am preparing myself for the inevitable coming of a new year, and therefore a new start. Work won't come in another five days, and so won't school, and it won't be until then that I'll be submitting a working progress (initial writeup and analysis of initial results of my thesis). For now, or in the near future (i.e. tomorrow), I'll be preparing myself for a marathon of Neverwinter Nights as I try to finish the second chapter and prepare to finish the game before buying the first expansion.

My thesis on a projective load balancing algorithm for a Beowulf cluster is taking a turn for the better (I hope) now that I'll be able to work on the cluster from home and implement the algorithm over the break. I am currently analyzing the results of the control setups and identifying certain trends and quirks (if there are any identifiable ones) from the data. I will be coming up with an online report of it, somewhere maybe in my personal w…

Missing in Action Still

I have recently been under detox (from the internet). Detoxifying myself of the highs that the internet brings to people like me, who like having to get the news at the touch of his fingertips, and communicating non-humanely with humans through the internet (IRC).

However, I do have nothing better to say, but that I am so tired that I want to do something already.

Kept chilled...

Progress Report

Computer Science 127
File Processing and Database System
First Semester, School Year 2002 - 2003


Progress Report



Group Name:
Name:
Section:
Date:



I. Group - any accomplishments or activities that involves the whole group, e.g. meeting, brain storming

II. Client - if you have a client, write here any transaction or consultation you did with the client.

III. Personal Achievements - write here the things you did in connection with the project e.g. you design interface, you created the table, write program

IV. Comments and Suggestions - any comments or suggestions regarding your group mates, project or client or anything you want to say with regards to the development of the project

V. Goals for next week - the things you plan to do for next week, this will serve as a reminder for you and the basis for the next progress report, if indeed you have a progress.

CMSC 172 - Project 1

CMSC 172 - Project 1

CMSC 172-Robot Modeling
Laboratory Project 1
Objectives

1. To write a solution to the gross motion path-finding problem in 2D space;
2. To initially write a solution for the above problem restricting only the mobile part to move in two translational dimensions.
3. To extend the code for the second objective above to include the third rotational dimension.

Prerequisites

1. Knowledge in C or Perl or Java programming.
2. Knowledge in scripting documents via the Hypertext Markup Language (HTML).

Required Output

You must publish a web page that will show the code specified in the objectives. The web page must be named ~your_login/www/cmsc/172/proj1.html. It must show your name, the specifics of your solutions and a link to the source code. All of your source codes must be properly documented. If you used multiple files for the source code, you must also provide a link for these. A Missing section must be included if not all requirements are satisfied ex…

Dynamic, Data Dependent Load Balancing for European Metacomputing

Dynamic, Data Dependent Load Balancing for European Metacomputing

A very important find, especially with regards to the thesis I am doing. I should therefore look into HiCon and give it a shot for setting up a link between the UPLB cluster as well as the ASTI cluster. Maybe even the other clusters from other Universities might want to give HiCon a chance, much like how the Europeans have done it.

Quoted from the site:

This paper presents the trans European Metacomputing research by the E=MC2 project. It illustrates the network, the trials and the experiences with extensive measurements between several European high performance computing centers coupled by ATM broadband networks, and investigates the possibilities and challenges of European Meta Computing on broadband networks.

For several trials the dynamic load balancing environment HiCon was installed on the clusters and several complex, parallelized applications were executed and observed on it. Additionally, several of these appli…

Motivation

It hasn't occurred to me that the only motivation I can come up with is the motivation that is not external -- but motivation that comes from within. If I hadn't been offered and given a job, I wouldn't have been turning my life around so drastically now. I'm getting up early, doing things more snappily, getting more professional (dressing up for everyday), and getting more from life than what I thought I could already get.

The thought of waking up to another day filled with duties that get rewarded seems a better incentive than the sheer love of doing it. We have got to get rewarded by other people more than just occassionally, but we need to be able to get motivation from ourselves. And what's better than doing your job and loving it? I couldn't call it work anymore because I enjoy doing it. However, it might be more appropriately called a profession because I want to excel in doing it -- and master things I should have mastered a long time ago.

Now I have my…

Instant Job

Right now, I am in the middle of completing the requirements for formally being employed by the Municipality of Calauan to administer their Linux based systems, and (hopefully) migrate the existing systems to Linux. I got the job offer just yesterday, and I've already started by installing Red Hat Linux 9 on a server (powerful P4 based computer) and am currently in the process of configuring the network and other services that is needed for the smooth operation of the systems being set up.

I will NOT in any means divulge the terms of which I will agree to work with, but getting paid with linux is like a dream come true. Although I cannot expect too much from the government (especially here in the Philippines) I owe much of the opportunity to my dear uncle Mayor. Without him as mayor, I will not be working right now.

The position is flexible and would allow me to study and do my thesis and work at the same time. I can work long hours (as long as I like) until I get the systems up, …

Missing in Action

I have been scouring the library for related literature, as I stumble trying to come up with papers that suggest that something similar to what I am doing would be feasible. I am also looking for papers that suggest formal proofs of the concepts they are using, and proposing.

I have come up with a few papers regarding load balancing, but most of them rely on looking ahead (or optimistic load balancing) of time, instead of using the current/previous performace of the system. I'm having problems coming up with literature that supports what I'm doing, and it's really not easy. I do have my best friend google and everyone's tambayan (the library) but the time it takes to actually look around and read through all the material is really just overwhelming.

I guess I just have to stay longer offline, and just take a peek once in a while.

Chillin'...

First day

It's the first laboratory session for CMSC 172, Robot Modeling and we are faced with a real world problem that needs a real world computational solution.

The problem is called the gross motion path-finding problem, and what it's all about is planning the route of a robot in 2 dimensional space, given the starting point, and solid obstacles in the defined space. It's like having a map of the city, and planning your route from one place to the other by avoiding obstacles (buildings, parked cars, passing trains, etc.) and getting the shortest path possible.

It's a machine problem which may be solved using C (which i prefer most) or Java which takes in the definition of the 2 dimensional space, the starting position of the "robot" and the target position of the "robot", and the definition of the solid obstacles in the two dimensional space.

I've heard a couple of approaches from sir Resty, and I'm stil scouring the web for possible hints as to h…

I feel the Power.

After a couple of hours setting up the cluster, I finally have it running. I finally setup my first SPARC based cluster. And I finally run the prime number finder. And believe me, I felt the power.

Even though the performance (in actual time units) wasn't better than the ASTI i686 based cluster, the fact that I was able to search through more numbers (1-9999999 compared to just 1-999999) is really something. I guess that's where the 64-bit architecture kicks in.

Aside from that, I can't really complain about anything else because everything I have been using at the ICS is provided for, and the things I am doing are for the benefit of the institute too.

The power, is definitely there -- and I have a feeling that it's going to stay there and become more useful as more and more people get to use the cluster. Someday I hope I'll be working with a firm that provides HPC solutions, because I definitely could get the hang of working on something I am passionate about. It …