A little background: my latest experiments are done on a C2Q Q6600 and on an i7 2600K (two and two). These are machines with 8/16GB of RAM and fast hard drives/SSDs. The speed compared to a T42 or the T60/T61 I use onboard is greater by orders of magnitude, not coefficients: I am used to using the scroll wheel
and getting instantaneous response from any program (be it Tradestation, Photoshop, what have you). And indeed, 1.3.6 and 2.4b without borders works that way. 2.4b with borders lags even when I drag the map, if the scale isn't on the step past the 150kish mark. Tapping +/- gives me a small delay which becomes large once the scale passes the 500k mark and there are LOTS of borders involved (testing mainly on the northern adriatic
sea, when you go 500k->1M you have all of dalmatia on the map, from Trieste to Dubrovnik). Large delay means multiple seconds, BTW. Dragging is all but impossible, as in the map will move seconds after your mouse motion. Then, once the scale gets up to 5M (about half the med on a 1600x1200 screen) the borders become fewer and scrolling is OK again.
VSS is at 222-224MB, private bytes around 119MB. Nothing big. Stuff running in the background doesn't really change things much (and it shouldn't, the machines are specced not to let that happen). I do not, in any case, run a virus scanner and the background defragger is paused during the workday.
The problem is pretty evidently with the part of the software
which either computes or renders the chart borders: I think developers should take a look. I'd love to but my copious free time (TM) is at historic lows.