home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.misc
- Path: sparky!uunet!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!jvnc.net!princeton!newross!elan!mg
- From: mg@elan (Michael Golan)
- Subject: Re: Is HPFS faster?
- Message-ID: <mg.711700324@elan>
- Sender: news@newross.Princeton.EDU (USENET News System)
- Organization: Princeton University, Dept. of Computer Science
- References: <SAREL.92Jul11192258@ford.ee.up.ac.za> <1992Jul11.192751.22613@unixg.ubc.ca> <1992Jul13.115126.26346@doug.cae.wisc.edu> <17JUL199210430534@cl2.cl.uh.edu>
- Date: 21 Jul 92 06:32:04 GMT
- Lines: 31
-
- csci11ba@cl2.cl.uh.edu (451924952 NEWMAN, DANA FRANK) writes:
-
-
- >I dont think you guys are approaching the problem the right way. I wouldnt
- >even consider the caches. Why not just turn both of them OFF and try it?
-
- Because this is not a THEORY problem. No one runs his system without some
- sort of cache, so who gives a damn about cache-off performance?!?!?
-
- >The FAT file system uses a linked list to find files.
- >The HPFS file system uses a B-TREE to find files.
-
- You mean to find blocks, of course, not files.
-
- >(For those of you who havent gone to "computer school" my apologies.)
-
- I would like to apologies on your behalf to these people. You prove why
- so many think that CS people are snobs, or can't do work in the real world.
-
- >So, if you look at the problem in the context of a linked list or a B-TREE to
- >do search operations, its pretty $%^&ed obvious which is faster!
-
- Oh really? How many performance tests have you done on this subject?
- Not a O(n) theoretical comparison, but a real one? B-trees (and other
- self-balancing data structures) are quite slow compared to simpler
- data structures, until you get to large sizes (and since each entry in FAT
- is 4k, you need very big files for B-trees to outperform FAT enough to
- show any performance difference in real life)
-
- Michael Golan
- mg@princeton.edu
-