home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Almathera Ten Pack 3: CDPD 3
/
Almathera Ten on Ten - Disc 3: CDPD3.iso
/
scope
/
201-220
/
scopedisk209
/
textinfo
/
report.doc
< prev
next >
Wrap
Text File
|
1995-03-19
|
7KB
|
175 lines
Archiver Report - By Michael Glew
~~~~~~~~~~~~~~~
I have been meaning to do this for quite a while, but have just not
had time... But now it's done! These tests were done on the following
configuration (Not that it matters - the results should be considered
relative to the other results):
Amiga 500
KCS Power PC Board (I'm using the Recoverable RAM drive for Storage
during these tests)
External Drive
Stock MC68000, OFS and OCS
I had each of the following archivers in my RRAM drive along with the
test subject file renamed to "tst" - I proceeded to create archives called
"a" (eg. a.lzh, a.zoo, a.lha, a.arc, etc). As zoo has 2 modes and only one
profix, I completed testing all but that mode before doing the tests for
advanced zoo seperately. All times are clocked using stop watch function
on my wristwatch. (...I don't happen to have an olympic standard stopwatch
just laying around, do I...):
Arc = AmigArc v0.23 03/04/87
Zoo = AmigaZoo 2.1 Beta 3 01/08/91
LZ = LZ v1.92 1991
/---------------------------------------------------------\
| Filetype : Executable |
| Size : 12248 |
>==========v==============v========v==========v=====v=====<
| Archiver | Data | File | Overhead | Add | Ext |
>==========+========v=====+========+==========+=====+=====<
| Arc a | 10170 | 16% | 10201 | 31 | 15s | 8s |
| LZ a | 8440 | 31% | 8468 | 28 | 10s | 6s |
| LZ -2 a | 8341 | 31% | 8369 | 28 | 10s | 5s |
| Zoo a | 10016 | 18% | 10184 | 168 | 10s | 6s |
| Zoo ah | 8344 | 31% | 8513 | 169 | 13s | 6s |
\----------^--------^-----^--------^----------^-----^-----/
/---------------------------------------------------------\
| Filetype : Crunched |
| Size : 13440 |
>==========v==============v========v==========v=====v=====<
| Archiver | Data | File | Overhead | Add | Ext |
>==========+========v=====+========+==========+=====+=====<
| Arc a | 13440 | 0% | 13471 | 31 | 21s | 7s |
| LZ a | 13095 | 2% | 13123 | 28 | 12s | 9s |
| LZ -2 a | 13029 | 3% | 13057 | 28 | 10s | 5s |
| Zoo a | 13440 | 0% | 17926 | 4486 | 16s | 4s |
| Zoo ah | 13031 | 3% | 13200 | 169 | 13s | 6s |
\----------^--------^-----^--------^----------^-----^-----/
/---------------------------------------------------------\
| Filetype : Text |
| Size : 15863 |
>==========v==============v========v==========v=====v=====<
| Archiver | Data | File | Overhead | Add | Ext |
>==========+========v=====+========+==========+=====+=====<
| Arc a | 6959 | 56% | 6987 | 28 | 12s | 8s |
| LZ a | 5165 | 67% | 5193 | 28 | 10s | 5s |
| LZ -2 a | 4855 | 69% | 4885 | 30 | 14s | 4s |
| Zoo a | 7076 | 55% | 7245 | 169 | 9s | 6s |
| Zoo ah | 4858 | 69% | 5027 | 169 | 15s | 5s |
\----------^--------^-----^--------^----------^-----^-----/
/---------------------------------------------------------\
| Filetype : Repeditive |
| Size : 102400 |
>==========v==============v========v==========v=====v=====<
| Archiver | Data | File | Overhead | Add | Ext |
>==========+========v=====+========+==========+=====+=====<
| Arc a | 97 | 99% | 128 | 31 | 18s | 23s |
| LZ a | 2175 | 97% | 2203 | 28 | 21s | 10s |
| LZ -2 a | 61 | 99% | 89 | 28 | 31s | 10s |
| Zoo a | 537 | 99% | 706 | 169 | 21s | 20s |
| Zoo ah | 63 | 99% | 232 | 169 | 36s | 14s |
\----------^--------^-----^--------^----------^-----^-----/
That's the data ... now here's the observations!!!
Observations
~~~~~~~~~~~~
/----------------------\
| |
| Minimum Output |
| Maximum Compression |
| |
>---v----------v-------<
| 1 | LZ -2 a | 100 |
| 2 | Zoo ah | 95 |
| 3 | LZ a | 87 |
| 4 | Arc a | 53 |
| 5 | Zoo a | 0 |
\---^----------^-------/
Compression Rating = Average of k*{1-[(Size-File/Size)-1]}
From this we can see that the new LZ DCA (Data Compression Algorithm)
has a greater compression ratio than before. This and the minimal overhead
that LZ/LZH has been known for has boosted it up above that of the new Zoo
DCA whose Overhead is still high. As you can see, the old DCA's are left
far behind in the stakes for high compression ratios.
/-----------------------\
| |
| Minimum Time |
| Maximum Efficiency |
| |
>---v----------v--------<
| 1 | LZ a | 100 |
| 2 | LZ -2 a | 79 |
| 3 | Zoo a | 69 |
| 4 | Zoo ah | 14 |
| 5 | Arc a | 0 |
\---^----------^--------/
Time Rating = k*[11.2 - (Average of Adding times and Extracting times)]
The older DCA in LZ seems to be faster than the newer one. Zoo and
Arc havn't the speed of the new LZ package. This comparison pushes Arc
back down in to the depths of computer history as far as efficient
Archiving is concerned.
/-----------------------\
| |
| * * * Overall * * * |
| |
>---v----------v--------<
| 1 | LZ a | 100 |
| 2 | LZ -2 a | 93 |
| 3 | Zoo ah | 42 |
| 4 | Zoo a | 12 |
| 5 | Arc a | 0 |
\---^----------^--------/
Time Rating = Average of Compression Rating and Time Rating
The LZ DCA's lead the pack over all with a choice of good speed or
good compression. There is little difference in the rating of 1st and 2nd
place. The Zoo's come a miserable 2nd with ratings half that of the
leaders. This shows that the Zoo DCA's are obsolete. Arc is Dead to the
world of Archiving so don't bother.
Overview
~~~~~~~~
If you wish to decide upon an Archiver, then I would whole heartedly
reccommend on of the Lharc/LZ archivers. LZ seems to be faster, and Lharc
does not support the -2 option for the Lempel-Ziv-5 Algorithm (LZ5). LZ is
by Jonathan Forbes as is available in the Public Domain. I would
reccommend paying for the registered version, as I believe the PD version
is Crippleware as far as the LZ5 Add goes - also it will not extract LZ1
files larger than 50Kb or so, so keep a version of Lharc handy if you don't
want to contribute.
I add this report to the public domain, and it may be freely
distributed as long as not profit is gained from it's distribution in any
way. This report is Copyright (Michael Glew) 1991 and may not be edited
and distributed in any way, spare that of typo' corrections....
If you wish to contribute (or are Mad and like trowing away money and
valuables) then you may send things (Australian Monetary Cheques, Amiga
Computers or hardware, CD Vouchers for Stores in my area, Letter bombs,
Virus Infested disks, whatever) to:
Mr Michael Glew
Lot 61 Warrambool Road
Tumbi Umbi
N.S.W. 2261
Australia
(Sorry not to include a Internet mail address, but Macquarie University is
stingy with their mail accounts - Take the hint if you're a member of the MU
Staff and are reading this)