home *** CD-ROM | disk | FTP | other *** search
-
- DISKTIME.PAS
- ------------
-
- Program DISKTIME is used to measure the average access time on the
- IBM-PC/XT and the IBM-PC/AT fixed disk. It is written in Turbo Pascal 3.0 as
- distributed by Borland International, 4585 Scotts Valley Drive, Scotts Valley,
- California, USA, 95066. This program will not destroy the data residing on
- your hard disk(s), and communicates solely through BIOS service interrupt 13h.
- Comparison times are from Table 5, Page 171, No 11, Vol 10 (Nov 85) of BYTE
- and are increased by 8.33 milliseconds to compensate for disk latency. If you
- have problems on a compatible, then it is not as compatible as you thought...
-
- Permission is freely given by the Author of this program to distribute
- copies of the UNMODIFIED program as you see fit. Please note that you will
- have to run ARC(hive) (included) to extract the sources from the library. Here
- is how you should proceed to generate a copy from sources
-
- N>arc e disktime - Extract all sources
- N>turbo
- TURBO PASCAL SYSTEM VERSION N.NN
-
- COPYRIGHT (C) 1983,1984,1985 BY BORLAND INC
-
- INCLUDE ERROR MESSAGES (Y/N)? y - Want diagnostics
-
- LOGGED DRIVE: N
-
- WORK FILE:
- MAIN FILE:
-
- EDIT COMPILE RUN SAVE
- DIR QUIT COMPILER OPTIONS
-
- >O - Option change
- COMPILE -> MEMORY
- COM-FILE
- CHN-FILE
-
- COMMAND LINE PARAMETER
- c - Compile to disk
- COMMAND LINE PARAMETER
- q - Main menu
-
- LOGGED DRIVE: N
-
- WORK FILE:
- MAIN FILE:
-
- EDIT COMPILE RUN SAVE
- DIR QUIT COMPILER OPTIONS
-
- >c - Compile program
- WORK FILE: disktime - Enter name
-
- COMPILING --> N:\DISKTIME.PAS
-
- LOGGED DRIVE: N
-
- WORK FILE:
- MAIN FILE:
-
- EDIT COMPILE RUN SAVE
- DIR QUIT COMPILER OPTIONS
-
- q - Exit Turbo Pascal
-
- N>
-
- The reason I wrote DISKTIME was to demonstrate how significant the access
- time is on a small computer system, and to encourage the development of
- a stand-alone utility which would compensate for this. Please note that
- MS-DOS allocates file space in "Clusters" of 2k blocks each, which are not
- necessarily consecutive. After you have been using your hard disk for
- a while, sequential file clusters will probably be scattered at random
- throughout your hard disk, and the time to load a file or program will
- therefore be approximately equal to (ACCESS TIME) * (NUMBER of CLUSTERS)
- For the 100 kilobyte Microsoft assembler, this works out to a potential
- of (73.3 milliseconds)*(50 clusters) or approximately four (4) seconds
- on the IBM-PC/AT !!! Hence the urgency in my appeal for de-fragmenters.
- Ideally such a disk optimization facility would be well-behaved, in that
- o It would be stand-alone, not requiring operator intervention
- o It would require no additional peripherals such as floppies or tape
- o It would not require scratch space on the disk being optimized
- o It would leave the disk in a usable state after abnormal termination
-
- Finally, I should like to thank the authors of the ARC(hive) librarian
-
- System Enhancement Associates
- 21 New Street
- Wayne, N.J.
- USA 07470
-
- and to note that it is easier to use, friendlier, and produces smaller files
- than the LU utility which the Turbo users group currently supports...
-
- Sincerely,
-
- Y.N. Miles
- c/o TRIUMF
- 4004 Wesbrook Mall
- Vancouver, B.C.
- Canada, V5Y 1Z7
-
- Ph (604) 222-1047
-