Circuit Broken

Deleting unused Linux kernels

bdsatish — Mon, 6 Feb 2012 19:12:39 +1100

A quick link to GitHub's kernel-clean script , authored by me.

kernel-clean is a shell-script to remove (clean / purge) unused Linux kernels from any Debian-based system (Ubuntu, Linux Mint, MEPIS, aptosid, etc.). It gets rid of deprecated kernel images, associated header files, dependent packages, symlinks and their associated drivers. Also, unpackaged cruft below /lib/modules/$KVERS will be deleted. Neither the currently-running, nor the newest kernels will be offered for removal. It is licensed under GPLv2 or later.

Shell script to build GCC Snapshot

bdsatish — Tue, 1 Mar 2011 21:06:05 +1100

I use GCC (gcc, g++, gfortran) and GDB on a daily basis. It's very important that the compiler itself be "bug-free". Most of the new features will be in GCC's SVN trunk; these will take many months or even years to be in the "stable" release. For most practical purposes, the trunk is "stable" enough to be usable -- you can try out the new features in the compiler and also get the latest bug fixes. I'm not a GCC developer myself but I do occasionally submit bug reports (like #44912, #44929 and others) and testcases. I do like to browse GCC source code, so I use the GCC weekly builds.

Here is small bash script. It downloads the GCC and GDB weekly snapshots from their FTP servers, unpacks the archive, configures and compiles (make && make install) automatically. I've put in some options for make that I normally use on my system. If you already have old source code and would like to update it to the latest weekly snapshot, pass the "--update" switch. I've put up the code on GitHub in case I update it later.

Language : : Google Transliteration IME for Sanskrit

bdsatish — Sun, 26 Dec 2010 06:25:37 +1100

Google Transliteration IME (Input Method Editor) is a free (gratis) software for many Indian languages. It lets you type Sanskrit, Kannada, etc. from your English keyboard. An excellent feature is automatic word-completion using its built-in dictionary (which works offline too). For example, just type 'kR' and it suggests कृष्ण, कृत्वा , etc.

It's annoying to note that there is no key mapping for Devanagari daṇḍa ( । ) or double daṇḍa ( ॥ ), even if you use the on-screen keyboard. Another problem with the Sanskrit IME is that it is not fully phonetic. For example, typing 'priyan' does not show प्रियन् at all! It suggests something else like पर्यन, प्रियान, etc.

I've solved both of these problems using the "Canonical Schemes". Basically, Google lets you define your own key-mappings if you are not happy with the default ones. It uses something called "classes" (not OOP!) in a scheme (.scm) file. I have created a scheme file which uses the Harvard Kyoto mappings for Devanagari transliteration. Download HarvardKyoto.scm here. Save it in C:\Program Files\Google\Google Sanskrit Input\Schemes and that's it. The next time you open Google IME, you'll see a 'C' icon. Just click it and start using.

The daṇḍa (। , U+0964)is mapped to vertical line ( |, U+007C). The file is simple enough to read, and you can modify it to ITRANS, ISO 15919, etc.

The file HarvardKyoto.scm is copyrighted by me (Satish BD) and released under the GNU General Public License v3.

Software : : Communications and Signal Processing

bdsatish — Sun, 15 Aug 2010 06:20:34 +1100

Happy Independence Day ! Celebrating 63 years of Indian Independence!

Of late, I've been doing a lot of programming in Fortran. No, it's not a dead language. Modern Fortran is a fully object-oriented language with array-like operations built into the language itself, making it more like "compiled Matlab". Fortran Wiki is an excellent starting point.

That said, I have written up a few thousand lines of code for Communications and Signal Processing routines. The project is licensed under GNU General Public License. So, it's free and open-source software. The project is hosted at BitBucket. I've tested it on GNU/Linux (Ubuntu) and Windows/Cygwin. It's nice to know that my software is equivalent to 40,000 US dollars of development effort. Wow ! Thanks to Ohloh for the metrics, I've added a badge on the left sidebar.

Language : : Writing Chinese in Devanagari

bdsatish — Thu, 3 Jun 2010 18:04:18 +1100

It has been six weeks in Hangzhou, China. This has given me enough time to decipher the Mandarin language. Mainland China uses Simplified Chinese. Chinese cannot be transliterated but only transcribed. What makes Chinese so different are the following facts:

There is no concept of vowels (a,e,i,o,u) or consontants. It's all about initials, finals and tones.
The same character can correspond to two (or more) different sounds. For example, 言 can be transcribed as yán or yín or yàn
The same sound can correspond to two (or more) different characters. For example, tā can be written as 他 or 她 or 踏

Here, I have come up with a scheme to pronounce Chinese as closely as possible by the Indian tongue. It's easy because Indian languages (say Hindi or Kannada) are "phonetic" -- we pronounce the way we write. One letter, one sound.

The rest of the article assumes that you have some knowledge of Pinyin, Devanagari, International Phonetic Alphabet, Vedic Accent, Sanskrit phonology and Mandarin phonology. Read on full article here !

Software : : Fixed-point Math for Embedded Systems, Part 2

bdsatish — Mon, 5 Apr 2010 16:49:24 +1100

As mentioned in my previous post , Fixed-point Arithmetic is proving more and more interesting. Ahem, I created an Open Source project at Bitbucket ! Here's the link : Fixed-Point Math Library. You can download the latest version of software for free and try it out. I've created versions 0.1, 0.2 and 0.3, but it's always suggested to use the 'TIP'

I managed to code a thousand lines in just three days (Happy Easter !) over the long weekend. So the following, features are complete:

Conversions among fixed-point, floating-point and ASCII strings (with or without rounding)
Extracting integer-part and fractional-part of Q-number, absolute value & saturation
Add, Subtract, Multiply, Divide (different versions for dealing with rounding and saturation in different Q-formats available)
Fixed-point Square-root and Inverse Sqrt

One more feature I added are unit-tests -- as and when I implement new code, CuTesT will automatically generate template for my test suite. Currently there are about 18 test suites, take a look at the test/
folder in the downloaded zip. Of course, all of them are passing !

Software : : Fixed-point Math for Embedded Systems, Part 1

bdsatish — Tue, 16 Mar 2010 18:28:33 +1100

Hi there ! As a DSP programmer, I often encounter this situation: I need to do a lot of math, but the underlying CPU doesn't have a floating-point unit at all ! So, the general technique used is Fixed-point Arithmetic. The basic idea is that every real number, like 3.14159, is mapped to an integer. We all know that integer arithmetic is faster than floating-point arithmetic. For ex, 32-bit integer multiply (MUL EDX) takes about 5 clock cycles on a 80x86, whereas 32-bit IEEE float multiply (FMUL) takes over 10 clock cycles.

There you are, I rolled up my sleeves and wrote my own C library for dealing with Q-formats. There is basic support for conversion from float-to-fixed, addition, saturation, etc. I will be updating it soon with algorithms for sqrt( ), sin( ), cos( ) and finally FFT. The current version of the software is just a header file. A sample test program is also attached. You can download the software here: FixedPointMathLibrary. The software is licensed under the Non-Profit Open Software License 3.0.

Language :: Catch up with News in Sanskrit

bdsatish — Sun, 31 Jan 2010 11:36:12 +1100

In this post, I'll introduce you to web resources that help you to catch up with the news in Sanskrit, free of cost, everyday.

1. All India Radio (AIR)
Since the 1980s, Indians are addicted to two things - the television and the radio, or more precisely, Doordarshan and Akashavani, as they were then called.If you think you are missing those good old times, here's some solace.The AIR has been broadcasting news in several languages everyday. You can subscribe to the RSS feed of Sanskrit news through this link:
http://sanskritdocuments.org/news/air_news.xml
I use Google Reader which neatly streams the MP3 right inside the browser. If youchoose to listen regularly, just register your name on this link. Of course, I am the 18-th on the list !You can find all previous archives and the news transcripts at Sanskrit Documents

2. Sudharma -- World's only Sanskrit Newspaper
Sudharma was started in 1970 by a Mysorean, Varadaraja Iyengar. As such, you'll find a lot of local news, in and around Bangalore.This newspaper is struggling for its existence. The least you can do is to spread awareness aboutthis to your friends. You can subscribe to its RSS feeds here:
http://feeds.feedburner.com/sanskritnews

Software :: One input, Multiple Outputs

bdsatish — Tue, 26 Jan 2010 20:34:50 +1100

There are various document formats in use today: Microsoft Word (*.doc), Open Office (*.odf), Portable Document Format (*.pdf), XML, HTML, Docbook, Rich Text Format (*.rtf), Unix man pages, GNU N/T/ROFF, GNU Info, GNU Texi, etc. However, the best document format that exists, according to me, is LaTeX. It has been in use for several decades and has stood the test of time. Nothing can beat LaTeX (or TeX, in general) when it comes to typesetting Mathematics. In this post, I'll give examples of a software converters which accept LaTeX input and convert it to various other output formats. I call this as the principle of "one input, multiple output" -- write once and use it in whatever format you want ! As an example, head over to my Channel Estimation article, which was automatically converted from LaTeX to XHTML.

Don't forget to subscribe to RSS feed! Read the full article!

Signal Proc :: Channel Estimation - I, Maximum Likelihood Estimate

bdsatish — Sun, 24 Jan 2010 20:50:31 +1100

In Signal Processing and Communications, one often has to deal with unknown signals. For example, the wireless channel (air medium) distorts the transmitted signal in so many ways (multipath fading, ISI, etc.) that the receiver has to use its own methods to recover the signal. Generally, such signals are modeled as stochastic random processes and statistical methods are used to "demystify" the received waveform. The air medium can be considered to be a time-varying FIR filter and the receiver unit must estimate the properties of this wireless medium.Such techniques are called Channel Estimation methods.

I have begun a series of articles, in which I'll explain the various methods of channel estimation. In this post, I'll explain a method called as Maximum Likelihood Estimate. I give an example on how to estimate the mean and variance of the wireless channel. To start with, I consider a time-invariant () system. Later, we'll consider time-varying systems. This article involves so much math that HTML can't cope with it. I used LaTeX and converted it to pdf. You can download it here.

Please comment if you have anything to ask !