home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.os.msdos.apps:4657 comp.bbs.waffle:2238
- Path: sparky!uunet!mcsun!sun4nl!alchemy!accucx!nevries
- From: nevries@accucx.cc.ruu.nl (Nico E de Vries)
- Newsgroups: comp.os.msdos.apps,comp.bbs.waffle
- Subject: Re: Stacker's TRUE compression ratio
- Message-ID: <3046@accucx.cc.ruu.nl>
- Date: 7 Sep 92 21:16:32 GMT
- References: <2TXmqB3w165w@infopls.chi.il.us>
- Followup-To: comp.os.msdos.apps
- Organization: Academic Computer Centre Utrecht
- Lines: 56
-
- In <2TXmqB3w165w@infopls.chi.il.us> andyross@infopls.chi.il.us (Andrew Rossmann) writes:
-
- > I've been using Stacker for my Waffle news spool for around a month or
- >so now. I had posted that at that time, SCHECK was showing a 1.4:1 ratio.
- >That has now increased to 2.4:1.
-
- > At first, this sounds nice, but I've been wondering just how good it's
- >really working. Due to the small size of the posts, I've been using 4K
- >clusters. My STACVOL is 60M. SCHECK shows 81,457,152 bytes allocated, but
- >stored in 34,096,128 bytes. Using 4DOS 4.01, I did 'DIR E:\ /S/U', which
- >gave me the total TRUE file size, and the total allocation size. This gave
- >me 39,677,349 bytes in files, allocated to 79,622,144 bytes. CHKDSK showed
- >1,830,912 bytes in 340 directories (which I presume are not compressed.)
-
- > If I take the total true size + what's allocated for directory entries,
- >I get 41,508,261 bytes. If I divide that by the 34,096,128 bytes that are
- >used in my STACVOL, I get a meager 1.22:1 ratio.
-
- This is tough to understand but I'll give it a shot.
-
- Stacker uses basically 2 techniques to save disk space. The first one is data
- compression (see comp.compression FAQ for more info on that subject), the
- second one is using smaller "microclusters" for storing the compressed file.
-
- The compression saves space by making the data need less disk space. The
- microclusters by reducing the slack (unused parts of clusters).
-
- The larger a file is the better it is to compress it. This has to
- do with the chance of repeated strings in the data which is one of
- the properties of data used by most data compressors. For large files
- therefore a better ratio is possible.
-
- For small files the microclusters will do a lot of usefull work. E.g. a
- 1 byte file can take up to 1:16th of its origional space.
-
- This also means your measurements are relative. If a lot of small files
- is checked for their compression ratio it will of cource be bad, nevertheless
- the saved space will be significant due to the microclusters.
-
- Now for the expected figures. Unlike most computer companies the
- gues at Stac, IIT, SuperStor etc are not excadurating. For an basic
- harddisk 1:2 compression is quite normal. For lots of small files,
- and/or large databases etc 1:4 compression is not uncommon. Does the
- harddisk contain lots of .ARC .ARJ .ZOO etc files (compressed ones)
- than the ratio will get lower because compressed files can not
- be recompressed.
-
- >Andrew Rossmann | Sysop of Infoplus BBS, +1 708 537 0247
-
-
- Nico E. de Vries
- _ _
- O O USENET nevries@cc.ruu.nl FIDO 2:281/708.1 COMPUSERVE ^%#$*%^
- o This text reflects MY opinions, not that of my employer BITECH.
- \_/ This text is supplied 'AS IS', no waranties of any kind apply.
- Don't waste your time on complaining about my hopeless typostyle.
-