the forums at degreez.net

It is currently Sat Nov 25, 2017 3:23 am

All times are UTC - 7 hours [ DST ]




Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 203 posts ]  Go to page Previous  1 ... 9, 10, 11, 12, 13, 14  Next
Author Message
 Post subject:
PostPosted: Mon Aug 01, 2005 2:39 pm 
why the hell programmers are just ignoring this thread???
this is the major problem of bittornado!
i'm a programmer, bot don't know a thing about this program
and bother of learn from the sources to solve this by myself.
but i guess it's not that hard to do. I'm just wondering if bittornado
programmers are capable of doing their job....

hey??? do you listen me??? where are you??????????


Top
  
 
PostPosted: Mon Aug 01, 2005 3:27 pm 
TheSHAD0W wrote:
A lot of people have complained about memory usage in the newest release under Windows. The application itself isn't actively using the memory, but Windows still reserves a large amount of RAM for its operation. I've traced the problem down to two features.

The first feature is disk read caching. This is designed to reduce the amount of thrashing your hard drive has to do. In order to do so, it reads an entire piece, which can be from a sizeable chunk of a megabyte to a few megabytes. RAM usage is equal to the size of a piece times the number of simultaneous uploads the client is performing. This shouldn't be that bad, but Windows' memory allocation is unable to keep its buffer from fragmenting, which combined with the large data blocks causes the buffer to balloon to the point where it may consume all available memory; at which point Windows takes steps to defragment the buffer, which increases CPU usage.

Disabling disk buffering solves the problem for normal operation, but the problem is often still caused by the fast resume feature. In order to provide the convenience of fast-resume combined with the reliability of hash-checking, fast-resume performs what is called "lazy hash checking"; it checks file hashes as pieces or portions of pieces are first read off the disk. By necessity, an entire piece must be read in order to check it, which raises the specter of memory fragmentation again.

I'll be working on some strategies for splitting up the buffers containing the piece data into smaller ones, which would be better able to fit within a fragmented buffer rather than cause it to expand. This will, however, increase overhead somewhat, though hopefully not to a point where it's noticeable.


Top
  
 
 Post subject:
PostPosted: Thu Aug 04, 2005 7:37 am 
how old is that thing? one and half years?


Top
  
 
 Post subject:
PostPosted: Fri Aug 12, 2005 2:12 pm 
Does it honestly matter how old that is? Since the problem remains unfixed, it stands to reason that the original request holds true.

I average upload speeds of 37 kB/s and I'm seeding a torrent with 2MB pieces. According to my system monitor, available RAM decreases by 1% approximately every three minutes.


Top
  
 
 Post subject:
PostPosted: Fri Aug 12, 2005 2:13 pm 
Oh yeah, and here's a pic of the thingy
<img src="http://img.photobucket.com/albums/v370/TrentArms/wever.gif" alt="Image hosted by Photobucket.com">


Top
  
 
 Post subject:
PostPosted: Sat Aug 13, 2005 11:35 pm 
Offline

Joined: Sat Oct 30, 2004 8:05 pm
Posts: 49
In my attempts to try to diagnose the problem, this appears to be the most likely issue (at least as best as I can tell):

http://evanjones.ca/python-memory.html
(read that plus the update:)
http://evanjones.ca/python-memory-part2.html

In short, it looks like unless something is done in Python, this is going to remain a problem.


Top
 Profile  
 
 Post subject:
PostPosted: Sun Aug 14, 2005 8:31 am 
Tim, TheSHAD0W already said that.


Top
  
 
 Post subject:
PostPosted: Tue Sep 13, 2005 2:02 pm 
Anonymous wrote:
why the hell programmers are just ignoring this thread???
this is the major problem of bittornado!
i'm a programmer, bot don't know a thing about this program
and bother of learn from the sources to solve this by myself.
but i guess it's not that hard to do. I'm just wondering if bittornado
programmers are capable of doing their job....

hey??? do you listen me??? where are you??????????


Suffering from this problem myself, and reading the explanation shadow has I already know they're working on it.

But what really gives light to some of the depth of your own stupidity is the fact that you said you ARE a programmer yourself and have NO idea how to solve this problem.

...Here's a little nugget of info for you, precious, they're programmers too and also have no idea how to fix it. If you're not satisfied by their free service, take your business elsewhere. They're doing this out of choice, not out of your demand.


Top
  
 
 Post subject:
PostPosted: Fri Sep 16, 2005 12:05 pm 
New ver 0.3.13 causes them to stick in the task manager again when finishing.


Top
  
 
 Post subject:
PostPosted: Sat Sep 24, 2005 9:45 am 
Yeah, this has almost been resolved in version T-0.3.12, now it is introduced back in version T-0.3.13 with full power :shock:


Top
  
 
 Post subject:
PostPosted: Mon Sep 26, 2005 10:00 pm 
don't know if this helps the authors of the program... but just wanted to put up some stats that may help in diagnosing/solving the memory leakage issue.


Laptop: Windows XP SP2, Intel P4 2.4 Ghz, 448 mb ram

Physical Memory - before system crashed
Total 457712
Available 27200
System Cache 98600

Physical Memory - right at system crashed ( i kept my eye on it)
Total 457712
Available 94000
System Cache 220000

I restarted a download as soon as I turned on my laptop. It downloaded at an average of 125 kbs, for 40 mins before it crashed.

hope that helps with something.


Top
  
 
 Post subject:
PostPosted: Tue Sep 27, 2005 9:17 am 
same guy as previous post.

i am using a laptop now, and am constantly having trouble with the memory leakage, but not too long ago, i was using a desktop PC, with less RAM and a slower processor. Yet i never had a problem with it crashing on me. Given, my speeds weren't as fast, but i was able to download for hours and hours and nothing my desktop wouldn't crash. On this laptop, if i download for more than an hour, it definitely will crash.

Why is that?


Also.... I have heard many good things about BitLord, and that there aren't many problems with it.... how do you people feel about BitLord? Any people who have seriously given it a chance?

This problem of memory leakage is detrimental to the application and its users; i'm looking into finding another program soon. This problem has to be addressed soon


Top
  
 
 Post subject: ram
PostPosted: Fri Sep 30, 2005 7:43 pm 
HI I NOTICED ALOT OF RAM USAGE
AND I GOT A TOO L AWHIE BACK FROM MCAFFE CALLED P C BOOSTER IT HAS ABAR ON SCEEN TO FREE RAM TAKES 30 SECS WHEN EVER YOU RUN[/b]


Top
  
 
 Post subject:
PostPosted: Tue Oct 04, 2005 10:14 am 
Image

Image

Image


Thanks.


Top
  
 
 Post subject:
PostPosted: Tue Oct 04, 2005 2:58 pm 
..wow, u got a helluva trojan over there. Didn't u notice anything?


Top
  
 
Display posts from previous:  Sort by  
Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 203 posts ]  Go to page Previous  1 ... 9, 10, 11, 12, 13, 14  Next

All times are UTC - 7 hours [ DST ]


Who is online

Users browsing this forum: Majestic-12 [Bot] and 9 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group