vaughan@orion.UUCP (Robert Vaughan) (09/26/86)
I recently had an interest experience in using the three systems above (MS-Pascal, TurboPascal, and MS-C). I wrote a filter program that did some character level processing on a text file, before outputting it (through the Tangent Technolgies PC Macbridge, usual disclaimers...) to an Apple Laserwriter. The program basically performed a state-machine processing and replaced certain macros with expansions that were translated into things like headers for memos, etc., on the LaserWriter. Having started with only MS-Pascal, I then recompiled using TurboPascal. After acquiring a copy of MS-C, i rewrote the code in C. The intersting part of all this is the resulting size and execution times, listed below: Data file size: 9744 bytes MS-Pascal executable: 25K bytes TurboPascal executable: 25K bytes MS-C executable: 10K bytes Time to execute: MS-Pascal TurboPascal MS-C --------- ----------- ------- 2:55 4:37 1:35 (time in min:secs, includes access to network and printing) This example is by no means a definite becnhmarking; and yet, for a first time user of the various compilers on the IBM-PC, it seems that the various claims of Turbo (fast faster fastest!!!) I have seen in some magazines are somewhat overrated. The character level I/O seems to be most efficient in C, and of the two Pascals, Turbo is definitely too slow.... It is nice to have a fast compiler and integrated editor package, but if the code can't hack it, I'll take the slow way during compiles every time! (disclaimers galore: I don't work for any of these companies, and if I did they'd probably shoot me for sending mail like this...)
smvorkoetter@watmum.UUCP (Stefan M. Vorkoetter) (09/30/86)
Without seeing the code, it is difficult to compare these. What kind of file did you use in the Turbo code. Turbo Pascal is incredibly fast at I/O depending on how you implement it. It would be interesting to see the code for these filters (at least the I/O parts)