try leetIRC on your 8088 now :)

[Mike Chambers]

i linked in a couple ASM routines related to direct location-to-location memory copies. the speed is great now i think. i can probably still speed up color parsing, but go ahead and give this first test compile a run on your 4.77 mhz 8088's and let me know how it works for you!

you can really see the screen write speed by using PgUp and PgDn to scroll the buffer. (btw i cut the backlog at 50 lines in this version, but when i speed up my AddBuffer i might bump that back up to 250)

thanks.

http://rubbermallet.org/LEETTEST.ZIP

 

[Trixter]

MASSIVE improvement! It's now much more usable on 8088. Printing/scrolling performance still isn't perfect but I'd say any 7MHz machine or faster would be perfectly happy with it.

 

[Mike Chambers]

MASSIVE improvement! It's now much more usable on 8088. Printing/scrolling performance still isn't perfect but I'd say any 7MHz machine or faster would be perfectly happy with it.

i have no complaints even on a 4.77 personally, other than when you first log on to a server or other times where you get a big flood of text from the server. after that, it's pretty quick during the normal course of chatting even in a busy room.

but yes, i definitely would like to see if i can speed up the new line adding at least 3x or 4x... the problem is not the buffer scrolling, that alone is nearly instant.

it's (obviously) just a little chunk of code like:

FOR n = 2 to buffersize%
  Buffer(n - 1) = Buffer(n)
NEXT n

i probably didn't need to type that to an experienced programmer like you, but that is quite fast the delay is hardly noticable if at all there. the speed issue lies in the fact that i have to parse each line and add an attribute byte for each character on every line. even the color code parsing doesn't even really slow it down. it's just the huge amount of string operations performed for attributes...

i don't know what i'm really going to be able to do about that because i really love having full color. i suppose a command line parameter to turn that off would be a good option if it's too bothersome on an old machine and you don't need/care about color.

 

[Trixter]

the speed issue lies in the fact that i have to parse each line and add an attribute byte for each character on every line.

Are you inserting attribute bytes into the same string with the characters? If so, don't do that -- build a new string, appending chars and attrs to the end of it. Appending to the end of a string is fast; inserting into the middle of a string is slow.

 

[Mike Chambers]

Are you inserting attribute bytes into the same string with the characters? If so, don't do that -- build a new string, appending chars and attrs to the end of it. Appending to the end of a string is fast; inserting into the middle of a string is slow.

actually i think i'm already doing it that way. it's been awhile since i've looked but i'm pretty positive. i'll need to pull up the code and check.