"Well i was just showing the complexity difference in strlens."

Heavy !!!

All of this ended up being a crash course about strlens. This is great. Come to find out I hace code in another program in relation to donkey i got from Mario. ( Sorry i mis-spell that name... He is one of those ultra serious coders that taught us all about strings. One of those guys that could never turn his nose up at you because of you engurance of asm. You know what i mean.

So i did some research.

I wanted to know where the "DE" was placed at.

It don't go to the data section... I goes to your cide section dead in where you did it proc or anything else.

Kool... So thats why it can't work with offset. It is a specal processer thing or something.

Anyway i replaced my catSTRING with you Spasm code and it work and i saw no difference.

I know about masm Strlen .. It read 3 bytes ahead and depening on what our code is and what api you are calling that read three byte ahead may mess something up at times.

That is were Jen strlen comes in... The Double J is always go but if something go wrong I now use the Single J of Jen Strlens.

I may not be a bit expert but i do test everything i ever did for weeks and months at a time in difference situations when i come up whit something new.

That how i founded out all of these little things that most will never see. I take my time because i was lucky to have that time to to dumb things with it....

I wish you would have your code tested just like they did with the others. I like it becase it is simple and is great for a beginnner to work with. Find how how far it lag behind in speed. I got a feeling it can't be behund by much.... What a 1/10 of a clock.

I'm doing it the long way for 1 or 2 byte now since i see the processer don't want to help.

I move it to an extra buffer

mov esi
mov edi
cmp
repsb

now i got it in the dummy buffer

than i cat it....

So now my .data? don't have a zero in certain part in that section ....

A waste of space but good enough until i learn a better way.

Will some one please clock and compare RobotBob last code above That is a nice of effort made with a boss attitute for asm.
Posted on 2003-06-30 03:28:56 by cmax
http://www.asmcommunity.net/board/index.php?topic=4058&perpage=15&pagenumber=4

Look at the graph..... I am lazy to play with profiler.

The code by iblis made used to scasb (apparently scasx is slow). :)

RobotBob's first code is probably the code by Agner Fog.
Posted on 2003-06-30 03:35:57 by roticv
That was an heck of a thread about strlen. It been some time It be nice to see lingo new set up in it to some day.
Posted on 2003-06-30 03:50:48 by cmax
Thanks for the link to this thread, i have never read it before.
Posted on 2003-06-30 03:56:17 by RobotBob
It was the greatest but you and donkey set the laws down. (What make strlen it tick)

Thanks
Posted on 2003-06-30 04:08:26 by cmax
The little one i posted , is leagues slower mainly because of:
--repne scasb--

Against jens:
FileSize: 52,200,000 bytes
Using hutch's tester.

Jens: 43-63 ms
The simple one in my post: 142.5 avg(10) with best time of 125.0

So it is the slowest of the list including the bscan.

So reconsidering... easy for begginers yet fast enough, the byte scanner wins.

But looking at the noted thread, sometimes it varies.
**it actually times just like ibilis first strlen post.**

That thread on strlen is just wonderful. After reading it I would like to compress all of the code and ideas into
a chm file or something. It would my a dandy strlen primer.


RobotBob
Posted on 2003-06-30 04:35:06 by RobotBob
If you do besure to add your and donkey explantion. Leave my dumb s**t out. :)

It is wonderful. Jen blue in than than blue out leaving us some heck of a dust.
Posted on 2003-06-30 05:00:38 by cmax