[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[bug-mdk] MOVE timing bug
From: |
Radu Butnaru |
Subject: |
[bug-mdk] MOVE timing bug |
Date: |
Tue, 11 Sep 2001 13:27:03 +0300 |
Hi,
I'd like to report what i think is a timing bug related to MOVE
instructions.
Although the info documentation reads that a MOVE instruction takes 1+F
execution times, where F is the number of blocks to be moved, and the
Knuth book says it takes 1+2F times (could be a translation problem,
though i don't think so), the mixvm seems to be timing it one execution
time.
My version of mixasm is 0.4.2. This is a sample file and its output:
* timebug.mixal
* Copies memory contents from 0-49 to 50-99
ORIG 1000
START ENT1 50
MOVE 0(50)
HLT
END START
$ mixvm -r timebug.mix
Program loaded. Start address: 1000
Running ...
... done
Elapsed time: 12 /Total program time: 12 (Total uptime: 12)
My expected output(last line):
Elapsed time: 62 /Total program time: 62 (Total uptime: 62)
or, acording to TAOCP:
Elapsed time: 112 /Total program time: 112 (Total uptime: 112)
I'm hoping this will be fixed soon, because it's annoying, especially
when you follow timings in order to increase speed.
Thanks,
Radu B.
______________________________________________________________________
Do you want a free e-mail for life ? Get it at http://www.email.ro/
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [bug-mdk] MOVE timing bug,
Radu Butnaru <=