I don't use a Windows, but from what I know Pd only uses one processor core. However, there are a couple of ways around this.
One way is to split up the patch, run each part of the patch with a separate instance of Pd and use netsend or OSC to communicate between the instances. I've not tried this myself, but check out this posting on the forum: http://puredata.hurleur.com/sujet-2957-cpu-load-only-use-single-core
The other way is to use the [pd~] object. This object was added to Pd 'to allow users to embed separate Pd processes inside each other, so that the OS can schedule the processes on separate CPUs' – check out Miller Puckette's paper, Multiprocessing in pd for more information. With the [pd~] object you are able to run different instances of pd through a master patch, but without the hassle of using OSC or netsend to communicate between them.