leisner@arisia.Xerox.COM (Marty Leisner) (01/25/89)
I recently wrote a COFF style libld implementation to unpack TI34010 coff files on ms/dos. Unpacking symbol tables caused lots of disk thrashing. I'm using Aztec C 4.1b. It seems the stdio routines reread the disk sector after every fseek. Fine. But it seems DOS goes to disk with every read. What does BUFFERS= do in config.sys? It doesn't seem to do much (DOS hits the disk every time it seems). Running with cache installed on a Compaq 386 indicates approximately 1 disk read per symbol. Doesn't DOS maintain some sorta crude buffer cache? Puzzled, Marty Leisner Xerox Corp. leisner.henr@xerox.com
nelson@sun.soe.clarkson.edu (Russ Nelson) (01/25/89)
DOS's buffering scheme is only applied to partial sector reads, which stdio is supposed to *prevent*. Sounds like you need your own disk cache. -- --russ (nelson@clutx [.bitnet | .clarkson.edu]) "I saved the whales!" - Rebecca L. Nelson, 3.5 years old, on receiving her Christmas present of a whale "adoption" certificate. Bless her liberal heart.
dixon@sagittarius.steinmetz (walt dixon) (01/25/89)
The dos disk caching scheme is not quite what one would expect from a
modern operating system. The cache is used for fat and directory blocks.
The only time the cache is used for normal file blocks is when a partial
block read/write is performed. DOS searches the cache sequentially
instead of hashing as well.
Walt Dixon {arpa: dixon@ge-crd.com }
{us mail: ge crd }
{ po box 8 }
{ schenectady, ny 12345 }
{phone: 518-387-5798 }
standard disclaimers applyralf@b.gp.cs.cmu.edu (Ralf Brown) (01/26/89)
In article <579@arisia.Xerox.COM> leisner@arisia.Xerox.COM (Marty Leisner) writes: }I'm using Aztec C 4.1b. It seems the stdio routines reread the disk sector }after every fseek. Fine. } }But it seems DOS goes to disk with every read. What does BUFFERS= do in }config.sys? It doesn't seem to do much (DOS hits the disk every time it seems). The buffers are used only for partial-sector reads and writes. If you read or write a complete sector, DOS transfers the data directly between the user buffer and the disk. Every MS-DOS "C" I've seen uses a buffer size of 512 bytes--exactly one disk sector. -- {harvard,uunet,ucbvax}!b.gp.cs.cmu.edu!ralf -=-=- AT&T: (412)268-3053 (school) ARPA: RALF@B.GP.CS.CMU.EDU |"Tolerance means excusing the mistakes others make. FIDO: Ralf Brown at 129/31 | Tact means not noticing them." --Arthur Schnitzler BITnet: RALF%B.GP.CS.CMU.EDU@CMUCCVMA -=-=- DISCLAIMER? I claimed something? --
mrwittma@phoenix.Princeton.EDU (Martin R. Wittmann) (01/30/89)
Obviously, disk caching is a problem--- surely some guru has come up with a reliable public domain/shareware disk caching program??? Can anyone enlighten me as to where I can find it? (Email if this is a stupid Q.) My PC at work has Lightning disk speed-up software, which works great, but I don't want to/can't afford to pay for it for my own PC!
nelson@sun.soe.clarkson.edu (Russ Nelson) (01/30/89)
Obviously, disk caching is a problem--- surely some guru has come up with a reliable public domain/shareware disk caching program??? An apparently freely copyable disk cache program is available for anonymous FTP from sun.soe.clarkson.edu, pub/pc/discache.arc. If anyone finds out that it isn't freely copyable, please email me. I have used it for years and it works for me. -- --russ (nelson@clutx [.bitnet | .clarkson.edu]) "I saved the whales!" - Rebecca L. Nelson, 3.5 years old, on receiving her Christmas present of a whale "adoption" certificate. Bless her liberal heart.
jborza%burgundy@Sun.COM (Jim_Borza) (01/31/89)
In article <5919@phoenix.Princeton.EDU>, mrwittma@phoenix.Princeton.EDU (Martin R. Wittmann) writes: > Obviously, disk caching is a problem--- surely some guru has come up > with a reliable public domain/shareware disk caching program??? Can > anyone enlighten me as to where I can find it? (Email if this is a > stupid Q.) My PC at work has Lightning disk speed-up software, which > works great, but I don't want to/can't afford to pay for it for my own > PC! PC Magazine published a utility named DCACHE.COM which can be downloaded from their Interactive Reader Service BBS in either source (.ASM) or binary (.COM). Their number is (212) 696-0360. 1200/2400 N,8,1 (Xmodem). I've used this Cache program and it's pretty good - a write-thru variety which can be configured for expanded memory (and extended, too, I think). It also seems pretty safe. The PC Mag. utilities are not PD but are avail- able free of charge. Source is nice to have, too. If you call them, look around for the docs for the software. At least you'll get a reference to the issue (vol/nmbr) and you can get it from the library. Jim Borza - Sun Microsystems Milpitas, CA Disclaimer? Sure, why not?
kiy@pte.UUCP (Kevin Young) (02/01/89)
In article <NELSON.89Jan30075823@sun.soe.clarkson.edu> nelson@clutx.clarkson.edu writes: >An apparently freely copyable disk cache program is available for >anonymous FTP from ... >-- >--russ (nelson@clutx [.bitnet | .clarkson.edu]) Would someone (russ?) please e-mail this cache program my way? We can't FTP and I could really use this utility. advTHANKSance! (ok, ok, so I stole it :-) Kevin -- >|< Kevin I Young uunet!edsews!pte!kiy Precise Technology & Electronics, Inc. Custom Automated Test and Measurement Equipment for Industry Clever Saying: "Fail to plan and you plan to fail" - Dr. Robert Scheuller
padgett@inuxd.UUCP (Gary Padgett) (02/03/89)
> In article <NELSON.89Jan30075823@sun.soe.clarkson.edu> nelson@clutx.clarkson.edu writes: > >An apparently freely copyable disk cache program is available for > >anonymous FTP from ... > >-- > >--russ (nelson@clutx [.bitnet | .clarkson.edu]) > > Would someone (russ?) please e-mail this cache program my way? We can't > FTP and I could really use this utility. > > >|< Kevin I Young uunet!edsews!pte!kiy Me too (please?) ----------------------- Gary Padgett AT&T Bell Laboratories Indianapolis, IN att!inuxd!padgett