melling@cs.psu.edu (Michael D Mellinger) (03/04/91)
I need to be able to generate delays of around 1/2 second. Anyone have a good way of doing this? I'm using MS C 6.0. The best I can come up with is a 1 second delay. Also, how does MS C compare with Borland C++ and Zortech C++? Borland and Zortech seem to be more aggressive at updating their compilers. Is ObjectVision good? It sounds like it might give Borland a definite edge over Zortech. I know this is a FAQ, but things seemed to have changed a lot recently in the PC C world. At least as far as the other guys are concerned. MS did release a bug fix, so everyone can safetly use 6.0 now :-). -Mike
tlglenn@cs.arizona.edu (Ted L. Glenn) (03/04/91)
One quick fix for generating delays is to fool around with loops. Doesn't
look good, but works. I can't remember, is there a sleep command in C?
--
-Ted L. Glenn "Don't worry, be happy!" <--Ack! Pffffhhht!
tlglenn@cs.arizona.edu
G19382105@ccit.arizona.edu G19382105@ARIZRVAX.BITNETmelling@cs.psu.edu (Michael D Mellinger) (03/04/91)
In article <1010@caslon.cs.arizona.edu> tlglenn@cs.arizona.edu (Ted L. Glenn) writes:
One quick fix for generating delays is to fool around with loops. Doesn't
look good, but works. I can't remember, is there a sleep command in C?
That's machine dependant. My program will run on an 8088 and a 80486.
-Mikevalley@uchicago (Doug Dougherty) (03/04/91)
melling@cs.psu.edu (Michael D Mellinger) writes: >I need to be able to generate delays of around 1/2 second. Anyone >have a good way of doing this? I'm using MS C 6.0. The best I can >come up with is a 1 second delay. When I need a delay, I just do something like this: (This is for the A86 assembler, and assumes a .RADIX 16, since all values are in hex) delay: xor ax,ax mov ds,ax les bx,[1C*4] mov [1C*4],offset int1c mov [1C*4+2],cs L1: cmp word cs:[ticks],9 ; 9 for 1/2 second delay jb L1 mov [1C*4],bx mov [1C*4+2],es mov ds,cs ret ticks: dw 0 int1c: inc word cs:[ticks] iret You oughta be able to get this to work in C somehow. Note that this isn't 100% right, since I don't chain to the original INT1C vector, but it is, as they say, good enough for government work...
joe@proto.com (Joe Huffman) (03/05/91)
melling@cs.psu.edu (Michael D Mellinger) writes: >I need to be able to generate delays of around 1/2 second. Anyone >have a good way of doing this? I'm using MS C 6.0. The best I can >come up with is a 1 second delay. Depending on your machine that may be the best you can do. For example the Hitachi 2020 (sold in Japan only) only generates (at least 4 years ago when I was looking at the problem) a clock tick once per second. For more convential machines you can use msleep(500) (millisecond sleep) if you were using the Zortech compiler. >Also, how does MS C compare with Borland C++ and Zortech C++? Borland If MSC doesn't have this function or handle C++ code then it's deficient in at least two areas. Other comments from me about MSC would be too biased to be worth your time reading. --- Zortech mailing list: send email to 'ztc-list-request@uunet.uu.net' with: Add: your-user-name@your-machine-name In the body of the message. --- Send Zortech bug reports to 'zortech-bugs@zortech.com' Send requests for educational discounts to 'zortech-ed@zortech.com' --- Zortech is my major source of income. Statements about them or their competitors cannot be totally without bias. -- joe@proto.com
sdawalt@valhalla.wright.edu (Shane Dawalt) (03/05/91)
From article <1010@caslon.cs.arizona.edu>, by tlglenn@cs.arizona.edu (Ted L. Glenn): > > One quick fix for generating delays is to fool around with loops. Doesn't > look good, but works. I can't remember, is there a sleep command in C? > The sleep command in Turbo C (C++) has 1 second resolution. Loop are not portable between machines due to different CPU speeds. About the only real fix is to take control of the timer tick interrupt. Shane(); -------------------------------------------------------------------------- From the keyboard of: email: sdawalt@cs.wright.edu Shane A. Dawalt --------------------------------------------------------------------------
mikel@teda.UUCP (Mikel Lechner) (03/05/91)
melling@cs.psu.edu (Michael D Mellinger) writes: > I need to be able to generate delays of around 1/2 second. Anyone > have a good way of doing this? I'm using MS C 6.0. The best I can > come up with is a 1 second delay. Here's some assembler for doing a 10-second delay using DOS's int 21h function 2ch call to retrieve the time. You may be able to modify this to use the hundredths field instead of the seconds field to get delays of fractions of a second. I haven't tried this. One caveat, Ray Duncan's book "Advanced MS DOS Programming" says: On most IBM PC-compatible systems, the real-time clock does not have the resolution of single hundredths of seconds. On such machines, the values returned by this function in register DL are discontinuous. I interpret this to mean the hundredths position does not increment by 1, but by a larger value. Therefore a simple test for equality will not suffice. I've included this snipit of assembler which may give you a starting point to solve this problem. ; ; This program scans the keyboard for 10 seconds looking for a typed ; key. If no key is pressed, exit is normal, otherwise, exit with ; non-zero exit code. ; ; Author: Mikel Lechner ; Date: 2/18/91 ; _text segment word public 'CODE' assume cs:_text,ds:_data,es:_data,ss:stack main proc far mov ax,_data ; Set up data segment registers mov ds,ax mov es,ax mov ah,2ch ; Get time of day int 21h add dh,10 ; Compute 10 secs into future cmp dh,60 ; modulo 60 jl noadj sub dh,60 noadj: mov when,dh again: mov ah,6 ; Check for pressed keyboard key mov dl,0ffh int 21h jnz gotit ; Yep, exit with code 1 mov ah,2ch ; Check for target time (seconds) int 21h cmp dh,when jne again ; Loop again mov al,0 jmp done gotit: mov al,1 done: mov ah,4ch int 21h main endp _text ends _data segment word public 'DATA' when db 0 _data ends stksiz equ 64 stack segment para stack 'STACK' db stksiz dup (?) stack ends end main -- Mikel Lechner UUCP: teda!mikel Teradyne EDA, Inc. 5155 Old Ironsides Drive | If you explain so clearly that nobody Santa Clara, Ca 95054 | can misunderstand, somebody will.
ahodgson@athena.mit.edu (Antony Hodgson) (03/06/91)
> The sleep command in Turbo C (C++) has 1 second resolution. Loop are not >portable between machines due to different CPU speeds. About the only real >fix is to take control of the timer tick interrupt. > I haven't been following this discussion, but could the delay() command (TC++) work for what the original poster wanted? It generates millisecond precision delays. Tony Hodgson ahodgson@hstbme.mit.edu
py@meadow.uucp (Peter Yeung) (03/07/91)
In article <sg2Gwuw?@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >I need to be able to generate delays of around 1/2 second. Anyone >have a good way of doing this? I'm using MS C 6.0. The best I can >come up with is a 1 second delay. > If it does not have to be really accurate, use a loop to check the clock counter at location 0x46c (4 bytes counter) until it advances 9 times. This method has the accuracy of within 1/18 of a second. A more accurate method is to write a counting loop to count how many times does it loop when the clock counter advance from x to x+1 and use the loop count as a timing base for a timing loop which take the same number of cycles to loop once as the initial counting loop. Yes, this implies intimate knowledge of the code generated by the compiler or write both loops in assembler (can be easily done in TC). >-Mike -- Peter Yeung Amdahl Canada Ltd., Software Development Center 2000 Argentia Road, Plaza 2, Suite 300 Mississauga, Ont. L5N 1V8 Phone: (416) 542-6300 Fax: (416) 858-2233
29195@ruve.runit.sintef.no (Gulbrandsen,Arnt) (03/11/91)
In article <1991Mar7.053927.12273@meadow.uucp>, py@meadow.uucp (Peter Yeung) writes... >A more accurate method is to write a counting loop to count how many times >does it loop when the clock counter advance from x to x+1 and use the >loop count as a timing base for a timing loop which take the same number >of cycles to loop once as the initial counting loop. Yes, this implies >intimate knowledge of the code generated by the compiler or write both >loops in assembler (can be easily done in TC). > NO! That won't work very well with multitaskers like DESQVIEW, which steal cycles but leave the system clock intact. Arnt Gulbrandsen University of Trondheim, Norway 29195@ruve.runit.sintef.no
eversole@acae037.cadence.com (Richard Eversole; x6239) (03/12/91)
--
Look into the manual at the _bios_* calls. I do not remember exactly which calls
you need but I think they are similar to bios_gettime. Also you could use the
interrupt routines to intercept the system clock every tick (18.2 per second).
I wrote some code that uses the bios call to do a programmable delay that has
resolution of less than one second. If I remeber, I'll bring it in from
home and post it here.
I was uses MSC 6.0a. The bios calls offer the advantage of working on
all 80x86 machines.
=====================================================================
eversole@cadence.com