Zugg Software :: View topic

Beginner Joined: 27 Oct 2010 Posts: 15

So..this may be a good place to start with what I am trying to get to. Tell me what things I need to show to figure out the root of the problem. What I have noticed is with my system I am getting different speeds (drastic) on different computers. I am using the CTRL + Q speed test (at least I thought it was a speed test) and I am noticing I get a speed of 24s or so. I think it is great for myself. I experience virtually no system lag and team fights are wonderful. However, on other people's computers with same exact system I use, they use the same CTRL + Q and are experiencing 50s+ speeds. For the most part I thought they were using better computers than I. I don't think it has to do with the OS because some of those guys use multi platforms on their comps and they get the same thing. Any idea what it could be? All CMUD 3.33a. My comp to start with is:
Microsoft XP Professional 2002 Service Pack 3
Dell D830
Intel Core 2 Duo CPU
T7500 @ 2.20 GHz
3 GB of RAM
**Edit Update: NVIDIA Quadro NVS 140M graphics card.

Of course no one else is about that uses my system at the second so I can't get theirs but is there any other information I need to find out? Is there something that will help get that speed down when the CMUD code is standardized across everyone's computer?

A) OK, edit:
First person 38.4s CMUD 3.33a
System: 3GB RAM
OS: Vista Home Premium, 32 bit
Intel Core 2 Duo CPU
T6500 @2.10 Ghz

Magician Joined: 28 Jan 2007 Posts: 430

umm, until now I thought that the smaller the number, the better? or am I misunderstanding something :/

Beginner Joined: 27 Oct 2010 Posts: 15

Apprentice Joined: 10 Apr 2009 Posts: 145

One factor, which you haven't listed in your specs, and which may be the big contributor to the differences, is the graphics card.

Beginner Joined: 27 Oct 2010 Posts: 15

Wizard Joined: 22 Mar 2007 Posts: 2320

As a side note--CTRL-Q is not a good measure of how fast Cmud will run. It may be a useful measure of some other things, though.

Posted: Mon Apr 11, 2011 4:56 pm

Any system running Vista or Windows 7 will have slower screen update times than a system running Windows XP. That is because of the Aero graphical "improvements" made to Vista and Window 7. To get the fastest screen scrolling speed in Vista and Win7 you need to go into your Performance settings for Windows and turn off the Desktop Composition feature. (In Win7 this is in the Control Panel, Performance Information and Tools, Adjust Visual Effects)

Note that turning off this setting will disable most of the "cool" visual features of Vista and Win7. But that's the trade-off: speed vs appearance. When Desktop Composition is turned on, writing to the screen from CMUD does not go directly to the screen anymore, but goes to an intermediate desktop buffer that Windows is controlling, and that is slower.

And yes, the speed of your graphics card will definitely have an effect on this, as well as ensuring that you have the latest drivers for your graphics card. It's not "3D game performance" that really matters, but more of the low-level 2D speed of the card and how well the drivers are optimized for 2D. Also I believe that a 64-bit system will be faster than a 32-bit system in this area, but I've never done a direct comparison.

Finally, as Rahab mentioned, keep in mind that this only effects your text scrolling speed. Things like your Triggers have a much greater impact on actual CMUD performance. So just increasing your text scrolling speed in Windows probably isn't worth doing if you also use lots of triggers anyway.

Apprentice Joined: 03 Feb 2007 Posts: 170

Wanderer Joined: 21 Aug 2008 Posts: 85

[edit][/edit]

Posted: Tue Apr 12, 2011 4:02 pm

Ctrl-Q dumps a large amount of text into the internal CMUD network buffer to simulate a lot of spam combat text. It then times how long it takes CMUD to process and display this text. So it's a test of scrolling speed along with trigger speed. This key function does not work when certain windows like the mapper are docked. So it is normally only used from a blank session screen.

Wanderer Joined: 21 Aug 2008 Posts: 85

[edit][/edit]

Posted: Wed Apr 13, 2011 4:03 pm

Yes, but turn on your triggers and you should start seeing a big speed difference.

Novice Joined: 10 Jul 2007 Posts: 47

OK, I got a little curious after reading this post and decided to crtl+q. I got 15-16s when I did it 4 times, yet I thought I had a slow scroll... when someone said that 24s was good. What is an "ideal" range?

Oh and I turned off the desktop settings thingy and it didn't change the speeds at all.

Posted: Wed Apr 20, 2011 4:28 pm

Sorcerer Joined: 15 Apr 2011 Posts: 809

Posted: Thu Apr 21, 2011 4:12 pm

Sounds like the MUD is sending the data in small packets. When set to zero, CMUD still refreshes the screen after every packet (to show MUD prompts that might be waiting for input). If your network connection or ISP is splitting up packets then you'll get that effect.

You can open the Script Debugger window and turn on the Raw Input/Output message to see the exact packets being sent from the MUD. The script debugger window will also show timing between lines and also show if some triggers or scripts are also running that might interfere. The Script Debugger is really the real way to track down speed issues. I use the Ctrl-Q function for internal debugging and it really isn't intended for normal use (which is why it's not documented)

Sorcerer Joined: 15 Apr 2011 Posts: 809

No, everything is being sent in one large packet (about 300 lines), and even if was being sent in several small packets, it wouldn't explain why there's a difference between a refresh rate of 0 and 1000... as you said, CMUD will refresh the screen after every packet anyway. With a refresh rate of 1000, the screen is clearly only being refreshed once, at the end of the packet. All I ever see is the last 30 lines or so, I don't even see the rest scrolling by. With a refresh rate of 0, I do see the entire output scrolling by. The screen is definitely being refreshed more than once (at the end of the packet), every few lines or so. I don't notice a difference between a refresh rate of 0 and a refresh rate of 5.

Posted: Fri Apr 22, 2011 4:48 pm

As I said, you need to look at the Script Debugger output to see what else is running. Or maybe you have some other software on your computer that is effecting how Windows is refreshing the screen. The MUD might also be sending GA/EOR markers in the packet which also get treated as end-of-packet markers (if the Use EOR/GA preference is enabled) which you would see in the raw input/output data in the debugger.

Sorcerer Joined: 15 Apr 2011 Posts: 809

I did look at the script debugger window, that's how I know everything is being sent in one packet. I think I'm just misunderstanding how refresh rate works. I'll start a new thread though, since we've gotten way off topic.

Newbie Joined: 29 Apr 2011 Posts: 6

I'm going to assume I am just very lucky in how it all plays out, I get times ranging from 4.1s to 5.6s, with a 50k line scroll buffer and the refresh amount set to 5.