home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!usc!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!yale!mintaka.lcs.mit.edu!bloom-picayune.mit.edu!athena.mit.edu!pshuang
- From: pshuang@athena.mit.edu (Ping-Shun Huang)
- Subject: Re: Why ZIP/UNZIP slower than DOS version?
- In-Reply-To: Ching-Hsiang Chen's message of Fri, 31 Jul 1992 05:09:21 GMT
- Message-ID: <PSHUANG.92Jul31024836@beeblebrox.mit.edu>
- Sender: news@athena.mit.edu (News system)
- Nntp-Posting-Host: beeblebrox.mit.edu
- Organization: Massachusetts Institute of Technology
- References: <1992Jul31.050921.19572@athena.mit.edu>
- Date: Fri, 31 Jul 1992 06:48:46 GMT
- Lines: 24
-
- In article <1992Jul31.050921.19572@athena.mit.edu> Ching-Hsiang Chen <chchen@stat.fsu.edu> writes:
-
- > I am using the zip/unzip binaries from tsx-11(or banjo?). The zipping
- > /unzipping speed is far slower than the pkzip/pkunzip DOS version,
- > especially for large files (>= 100k). Anyone has the same problem?
-
- I assume that the binaries you are using are results of the recent port
- to Linux of the code from the Info-ZIP project. Two reasons they seem
- slower than PKZIP/PKUNZIP 1.1 for DOS are: (a) PKZIP/PKUNZIP are coded
- in assembler (probably with reasonable care toward optimization by
- PKWare, Inc.) whereas Info-ZIP source code is portable C and not
- hand-optimized for any particular platform; (b) Info-ZIP's recent
- releases by default use a compression algorithm whimsically named
- deflation (as opposed to PKZIP 1.1, which uses implosion)... PKZIP 2.0
- will be using deflation as well, and my experience (playing around with
- PKZIP 1.93 alpha) is that deflation tends to be a bit slower than
- implosion anyway (but gives ratios in the ballpark of ARJ). Whether this
- (performance hit) is true in the general case or will be true for PKZIP
- 2.0 release, I don't know, but that the two archivers are using
- different algorithms is a good reason for their performance to differ.
-
- --
- Ping Huang (INTERNET: pshuang@athena.mit.edu), probably speaking for himself
-
-