Can this code run faster? - GameCreators Forum

Author

Message

Vespar

23

Years of Service

User Offline

Joined: 17th Jul 2003

Location: USA

Posted: 17th Jul 2003 17:35 Edited at: 17th Jul 2003 17:44

Link

+ Code Snippet

fs=file size("anyfile.txt")

for a =1 to fs
  byte= read byte from file ("anyfile.txt",a)
  if right$(bin$(byte),8) = "11111111" then goto exit_file_search
next a

exit_file_search:
print a
end

This code can take a very long time to run. I'm using
files as large as 5 meg to 10 meg. I would like this to
run in as little as 1 second. As it is now I've seen it take
as long as 30 seconds.

If it was converted to c++ how fast would it run? I dont
know too much about c++ and how to get DB to read c++ files.
Any help would be appreciated.

Back to top

Profile PM Email

Van B

Moderator

23

Years of Service

User Offline

Joined: 8th Oct 2002

Location: Sunnyvale

Posted: 17th Jul 2003 17:43

Link

You could load it as a memblock, I imagine memblock loading would be a great deal faster than loading byte by byte.

Also, try and avoid complex checks, instead of looking for 11111111, look for byte value 255.

Van-B

My cats breath smells of cat food.

Back to top

Profile PM Email

Shadow Robert

23

Years of Service

User Offline

Joined: 22nd Sep 2002

Location: Hertfordshire, England

Posted: 17th Jul 2003 18:22 Edited at: 17th Jul 2003 18:24

Link

+ Code Snippet

fs as dword
af as string
af = "anyfile.txt" 
fs = file size(af)
a as dword
by as byte

open to read 1,af
for a=1 to fs
  read byte 1,by
  if by=%11111111 then goto exit
next a
close file 1

exit:
  if file exist(1) then close file 1
  print str$(a)
  wait key
end

would probably run a but faster in C++, but its more the way your checking than anything else

Back to top

Profile PM Email Website

Fallout

23

Years of Service

User Offline

Joined: 1st Sep 2002

Location: Basingstoke, England

Posted: 17th Jul 2003 18:24

Link

I've never really used files in db, but Van is right in that you need the whole file in memory before you start reading from it, else it'll be dead slow. There must be a way in db to put all that data in to RAM before you read from it. If there is a make memblock from file command, then that has to be the way.

Machine: P4 2200, 1GB RAM, GeForce4 64MB, Audigy Platinum
http://www.breakbeat-terrorism.co.uk
(It's not all about the coding)

Back to top

Profile PM

IanM

Retired Moderator

23

Years of Service

User Offline

Joined: 11th Sep 2002

Location: In my moon base

Posted: 17th Jul 2003 20:05

Link

There is

+ Code Snippet

function LoadToMemblock(MB as integer, Filename as string)
   open to read 1, Filename
   make memblock from file MB, 1
   close file 1
endfunction

Back to top

Profile PM Email Website

Shadow Robert

23

Years of Service

User Offline

Joined: 22nd Sep 2002

Location: Hertfordshire, England

Posted: 17th Jul 2003 20:30

Link

for process loading it doesn't really matter speed wise using either the memblock or direct file access, they both load the file into buffer memory - the only reason you'd use the memblock is for continued used.

so ops outside of loading data, like a realtime textpad and such.

Back to top

Profile PM Email Website

IanM

Retired Moderator

23

Years of Service

User Offline

Joined: 11th Sep 2002

Location: In my moon base

Posted: 17th Jul 2003 20:51

Link

Hmmm, not quite true. I find it a lot faster to load the file in one go when I'm using byte operations, like when I did that height-map stuff (57MB file).

It's also easier - a 'for' loop with a known limit is easier to use (slightly) than detecting eof.

Back to top

Profile PM Email Website

Sorry your browser is not supported!

DarkBASIC Professional Discussion / Can this code run faster?