Is there any way to make KM to load up faster? I really do not want it to take up RAM by loading it at startup and having it stay there. I would like it to just load up faster when I want it? Heck even IE loads faster and other more bloated programs!!

I honestly think this is the main shortcoming of KM.

IE loads fast because all the libraries needed for it to function are *already* loaded into memory on windows startup. Ever wondered why your windows takes so long to start? Because all kinds of unneccessary crap is loaded with it:-P

Parts of IE are always loaded with windows even if you don't use IE. Thats why IE is undeniably fastest loading browser (save the old Netscape 4.x) on windows. Also, for example, Galeon starts on Gnome much faster than on KDE just because Gnome already has relevant libraries in memory whereas in KDE they have to be loaded.

So IE kind of cheats on this. It's not fair comparison. IE takes much more memory than KM ever will but you don't see it because it goes into common windows memory pool. If IE wouldn't be there then your windows would take up lot less memory when running. But there is nothing much you can do about it.

as far as making k'meleon load faster, disable any plugins you don't need, and keep your HD in order... obvious stuff... some of the stuff k'meleon needs at startup is kept in a compressed archive; i don't know if this is a factor, or whether anything could eventually be done about it in the process of individuating from mozilla...

all told, it's really not that bad, though.

Good point about IE and Windows, but if KMeleon wants to compete well with other browsers, especially IE, it has to load up much faster, because it is really that bad! And yes I defrag once every month, have nothing useless installed, no unnecessary plugins loaded and everything is in order - Performance is an issue for me as you may have noticed

As for loading speed, Mozilla 0.9.7 (the latest) does indeed load up quickly - almost comparable to IE6! So it just may be that KM will get a boost when it gets done with the latest Mozilla, which presumably is more optimised and has less debug code in it... Fingers crossed and I'll keep an eye on things.

Something's upside down here

On what system you have Mozilla starting up faster than KM? On both of my comps (p2 350 & p3 800) KM wins Moz hands down in starting speed and in general responsiveness. Are you absolutely sure that you don't use Mozilla loader? I use nightly Moz so I should pretty much get any possible speed improvement they've done since 0.9.7. There is no way Mozilla can be faster than KM

I tried to make a honest list of loadup speeds for the browsers I have:

Scope (using IE).....3 sec
Opera 6.0...............4 sec
IE 5.0 ....................7 sec
K-Meleon ..............11 sec
Mozilla ..................16 sec

I am using Win98SE on Celeron 433MHz with 192MB of memory. No loaders for browsers except for the built-in IE ;( Minimum set of plugins, if any, for each browser. So K-Meleon can be considered as relatively fast but I like it for other reasons. (I am not using K-Meleon loader because it is unstable on my otherwise quite stable computer.) I know nothing about Netscape, as I have never liked it. On Opera, the mail and news plugins are disabled; Mozilla loads without the composer.

Scope (w/ IE) .......... 7
IE 6 ................... 3/4
Opera 6 ................ 4
Moz .................... 12 maybe, hard to tell
km ..................... 7

on my celeron 566, 128 meg. Mozilla might be slightly faster, but not a lot. km is not spectacular at loading, but is easily quick enough to not be a problem.

I have the same system as Deft running IE6,Opera6 and K-Meleon and I will be hard pressed to tell the difference in loading speeds and I am not using the KM Loader.
What are we talking about 3 or 4 seconds ?,I would not even waste the time to check.Loading time is not an issue with KM as far as I am concerned.

I have done my own benchmarks using a stopwatch.

Set up:

PIII 600, 128MB RAM. Running Win2K SP2 with all the updates. Hard disk defraged recently (less than a week ago).

Loaded Programs:

Turned off everything except for AVG Antivirus. Also whatever Win2K loads up. And there is TClockEx also loaded up. There is no hard disk activity, and the system is just waiting.

Test Design:

Since IE has an advantage by being in memory already, I decided to level the playing field a bit by loading up each browser once and then benchmark the second loading time within 30 seconds of loading it the first time. Each browser is set to load up a local startup page I made for my self. The page contains nothing special - valid HTML 4.01 Transitional with an embeded stylesheet (using style tags in the head). No loaders were used for any browsers (except for the fact that 1/2 of IE is already loaded with Windows). During the second load for ALL of the browsers tested, there was no hard-disk activity, implying all of the loading came from a cached copy from RAM.

I think this strategy is the best I can achieve. Please recommend suggestions if you see a flaw.

Results (2nd Loading Time):

IE 6.0: 2 sec
Mozilla 0.9.7: 6 sec
KM 0.6: 9 sec
Opera 6.0: 3 sec.

nine seconds from memory?!!

something's definitely wrong. after the initial grinding start-up, if i close k-meleon and then start it again, it's more in line with the time you show for opera... maybe i'll do some clinical evaluations of my own.

ok, here's my own attempt to quantify what i've only felt intuitively up to now:

system is k62 450, 192MB, which is probably not terribly significant.

running win98SE stripped of most non-essential crap, with IE 5.5.

no mozilla or netscape installed... this may be significant? **

both home pages set to about:blank.

reboot windows, run IE: 7~8 seconds.

close IE, then run again: 1~2 seconds.

reboot windows, run KM: 12~13 seconds. (no loader)

close KM, then run again: 2~3 seconds.
this agrees pretty much with my casual assessment from daily use.

i present it not as 'proof' of anything, nor in support of any argument,but only for perspective; it explains one reason that opinions on loading speed vary -
obviously, the experience on which the opinion is based varies quite a bit itself.

** i should also note that i loaded the following plugins with k-meleon:

IE favorites plugin
macro extension plugin
toolbar control plugin
history plugin
plus a <a href="http://tln.lib.mi.us/~amutch/pro/kmeleon/khistfiles.htm">couple of other files</a> not included with the release.

...and that i didn't load any third party plugins, or the mozilla bookmarks plugin, since i don't have any bookmarks to load. this may be a factor?

Some more words about my test since it apparently has aroused some activity here.

I did not want to prove anything nor to support any argument. I think that K-Meleon's loading time is just normal and I am satisfied with that. It was just for reference and, possibly, to see the results of other users, if anybody else was to present similar data. I measured the times exactly, everything was for the first-time loading.

It is interesting that I may load and unload K-Meleon as many times as I like, but the result is still the same: 11 sec. So how's that fast loading K-Meleon from memory achieved? (I never installed the loader. However, I am using about 10 third-party plugins.)

well, one more thing that i can mention that may or may not have something to do with startup time: under system>file system properties i have my 'typical role of this computer:' set to 'network server' since that seems to work with memory cache more vs. disk loading... for whatever that's worth.

Andres, when data is read from disk, it is usually stored in a cache in memory. This way, any later access of that portion of the disk can come from memory, which is a few hundred times faster than the disk. So, if you load KM once, then unload it, all of the data it loaded from the disk will still be in memory. Reloading KM at that point *should* just grab it all from memory (you shouldn't see the hard disk light flash at all).

The best way to check is to close everything, wait for everything (hard disk activity) to settle down, run KM, wait, close KM, wait, and then run KM. If it doesn't load in under 5 seconds or if there is hard disk activity, something is wrong, because you have more than enough RAM. On my PII-266 w/ 96mb, it loads in 2 or 3 seconds from memory.

po, I actually just looked into that. <a href="http://support.microsoft.com/default.aspx?scid=kb;EN-US;q140679">Microsoft says</a> that this only relates to how much info it keeps about recently opened folders and files, and it doesn't seem to have anything to do with cache. I don't know of any way to actually configure the cache, at least not in Win9x. I think it's just an automatic thing that takes as much RAM as it can, generally. You can check the current size of your disk cache with sysmon.exe. Mine is currently about 24mb.

Mark,

I think it has to do with my own configuration. As I had much less memory once and Windows was quite leaky, I reconfigured some settings (using CacheMan optimization and Flexy for conservative swap file usage; there are some more settings for memory; that reduced hanging and illegal operations / page faults to practically zero and I like it this way since everything is working normally). So I think my loading time may really be an exception with this regard. (I did not figure that out at once.) Thank you, anyway, for your time and the information :-)

yeah, and thanks for clearing up the 'typical role' thing... it's just one of those things you're s'posed to do, but i never looked up the real significance. no wonder i never seemed to notice much of a difference...