TNT2 bêta

Hello

Since developing compression software, i've been asked sometimes what
made TNT the better compression ratio software available on Saturn-
based system.
In order to answer these questions, i decided to have a look at TNT
code.

"Having a look" is more easily said than done, as TNT binary code is
marred with many obfuscating techniques, especially Program Counter
manipulation, specifically designed to defeat decompilers & splitters.
Therefore, getting to the ML code is a little challenge in itself.

But once there, i was quite disappointed by what i discovered.
I don't know if this was already a well-known fact, but TNT is just a
rip of BZ, which source code is freely available. Same code, same
registers, same algorithm, everything is exactly at the same place,
including errors. Now of course there is a modification that explain
the ratio difference, but it is somewhat simple and well located : the
"encoding trigram" selected by Mika Heiskanen to code the "length
field" is replaced by a limited version of Elias-Gamma-Code, which is
more efficient, but also more computation intensive.

Now for the good part : if TNT is a rip with little modification from
BZ, then creating a TNT2 version starting from BZ2 can be quite
simple, and should result with approximately the same benefit (faster
compression/decompression, lower memory usage and higher compression
ratio).

Hence TNT2, the new "highest compression ratio" Saturn-based
compressor, for HP48,49,50, currently at Bêta2 stage, which in fact
should be called "Release Candidate" as it is already extensively
tested.

You can find it at following address :
http://HPUtils.Webhop.org
or
http://phantasie.tonempire.net/utilitaires-f7/tnt2-aka-bz2-with-elias-gamma-code-t77.htm#95
or (direct download)
http://img21.xooimage.com/files/7/c/2/tnt2-beta2-bb5dce.zip

As previously with BZ2, i'm willing to propose an opened beta phase
before actually declaring it "final release", although at this stage
TNT2 has no more any known bug.


Regards
.

Relevant Pages

  • Re: Gasoline grade BTUs per gallon?
    ... They have much higher compression ratios than gasoline ... built to use say, 87 octane gasoline, you will see no "preformance ... Higher compression ratio engines can make better use of it. ... If my engine requires higher octane fuel, ...
    (sci.energy)
  • Re: Why Cant A Fuel injected Petrol Engine be as Efficient as a Diesel?
    ... >compression ratios. ... This limits the compression ratio. ... diesel engines compress air only and add the fuel after that. ...
    (sci.energy)
  • Re: [fitsbits] Rice compression from the command line
    ... but beats gzip and bzip2 handsomely in compression ratio. ... Which raises the general question of constructing a figure of merit for data compression. ... Discussions like this usually focus on compression ratio, the speed to compress and the speed to decompress, but there are a number of important, less quantifiable, parameters: ... The FITS tile compression convention is one step toward greasing the rails. ...
    (sci.astro.fits)
  • Re: Compression filter for Loopback device
    ... The idea i had was to mount a swap partition over a compressed block ... > Anyway, my main concern was compression ratio, not performance. ... >> I'm very surprised you got ratios better than CramFS, ...
    (Linux-Kernel)
  • Re: New Compressor for HP 48,49&50
    ... I tested SM48 with different packers, ... *Failed* (Flash outputs an error when compression not better than ... TNT: 1123.5 B ... tiny nibbles gain. ...
    (comp.sys.hp48)
Loading