Does anyone actually know how to reduce the memory usage of iconX or is that just what I get? Seriously.
Kudos to the real skinners out there, how do you find the time?
Oh yeah I'm not interested in memos, dead horses or underpant comparisons ,no disrespect
that probably makes me a noob or something, I know.Maybe even a retro, oldskool noob
C'mon hit me, I'm ready to accept that 130,000k plus is my own doing. Is it due to the amount on the desktop at any given moment or something, or is that stoopid talk?
Yes I'm not a programmer.

I would say that its not typical IconX usage and would wait til a SD member replys as this is the weekend and they are off...

I would say IconX has a memory leak. Mine was at 8Mb at 11:00 today, At 12:30 it was 16Mb. Now, at nearly 16:00 it's 48Mb...

Mine uses a bit more than 500kb of memory. Granted I don't really care cause ive got lots of ram... (im guessing wbvista.exe = wbload.exe for xp)

Mine's at 1,104 K, but I'm on 32-bit not 64-bit like you.

I also have a VistaSrv "WindowBlinds Service" process taking up 672 K.

Not quite sure I understand this one but

It means its using 60k Harley   

John ,I know , but does that not seem quite low to you!

It's not a RAM problem that people suffer & is easily understandable as to why some not so informed people would come to the conclusion that customization software affects performance, on the contrary it has EVERYTHING to do with graphic card power,i should know i experianced it on my own machine that had a 32mb intergrated with 1gb of RAM & more than enough harrdrive space & speed running xp,as an example most blinds would affect say start menu mosue over load speed to get too the all programs list it generally will lag (it's possible that if tweakui were operating then said menu may appear sooner if slider was used to decrease the pause time)however unloading wb would sort it out or by using a minimal blind like liquid 1.1,OD & cxp did'nt appear to affect anything that i can remember,ip did'nt seem to affect anything but adding winfx forget it,as for iconx or widgets i can't comment as i never used to use them,adding a AGP 256 G card fixed any so called performance issues which i assume anyone complaining about memory hogging customization software is saying is in actual affect mistaken graphical card issues for RAM.

Thanx peoples, I am still looking at this,BTW here are some specs they could be better but will have to do for now.
Thanx scorpy, yeah I am totally NOT informed and I'm not looking to blame anyone for this memory issue, it is a weird one though from my myopic POV
Here's a GREEN wallpaper I made - I call it "goblin" feel free to use it, sorry it's not widescreen.WWW Link
[/link]

Chuckeye you watch far too many re-runs of kojak

Is it really needed to inject the code (not dramatic, code is shared by default) but also all the skin data and rederung objects ni all processes for rendering styled windows components? y opinion is that it just need to inject the basic interface that allows interacting correctly with the application, and for the rest the code can be a simple proxy client exchanging data with another process that effectively makes the actual rendering on screen; this just requires sharing the GDI window handles and making their surface or part of their surface accessible to the other process. The issue could be that there would be loss of performance coming from the need synchronization if the separate rendering process is single threaded and uses a serialization queue. However, the rendering process could have several shared worker threads for actually making the rendering. Note that not everything requires rendering in a separate proces, there are simple thnigs that can use shared atoms for things like suctom brushes, and there are lot of things that require just plain brushes for renderng retangles, text... This should not require synchronization in the renderer as this could still be returned to the proxy client as a list of GDI actions to perform for the rendering (think abuot the possibility of Windows Meta File formats for sending instructions to the client proxy). Also, an application can have many windows, but all of them are in the same virtual memory space: there should be a single thread for the whole application to perform the rendering and implement the proxy client. The serialization will be limited by using a process owned queue: this is the role of the windows message loop. Data sharing is really something to consider. Most of the time it's not really useful to maintain separate state varaible for each window used by an application (including the window created by GUI components) and a "light" window can fully emulate the function. This won't affect other processes or the Explorer process or the Desktop that are running in their own environment. The real issue is: can we figure out a system to use data sharing with a maximum of concurrent threads, and create a GUI renderer worker thread only when there's too much contention on the shared data? Generally the limitation is impossible to predict in code, without looking at the possibility of the display driver and hardware, that also implements its own rendering queues... Code sharing is a non-issue. Not using data sharing is the real issue, because it will increase the "working set" and so it will flush out the memory cache (in processors) or pages to disks, and the pages for code coudl be affected, or the CPU will need to add cycles to reread the missing memory in cache. Note that most memory needed should be allocated for the GPU, not the CPU. Beware about the unmonitored use of GPU ressources: if the GPU is overloadedn, the CPU will need to wait and this wil block the processes, that may appear idle: so even if the CPU usage is low, it does not mean that the application is running well, as the applications (and the skins renderer) may be blocked by GPU ressources becoming full, and then the appliation queue will also become full, taking lot of memory for its queues. Optimizing is not something as trivial as you think, and even your performance monitoring tool may be lying: please consider monitoring also the GPU resources!

Wow, way to hard to read. You may even had made some valid points. Try using control Ctrl m the next time to put in some breaks for paragraphs or to establish points of discussion. Not trying to an ahole, just offering some advice.  

My objectdock takes up around 4000K and when its being used, around 8000K