After doing some lengthy research into input methods for my mouse I started to implement WM_Input. The problem i am having is that: I'm running a single thread. My program renders at 60 frames per second and I'd like to keep it that way to avoid tearing. This leaves the windows callback stack to be checked 60 times per second. My mouse is running at 200hz. 10 seconds of mouse movement - gives 33 seconds of on screen movement I believe because the mouse messages have been queued up (I did the math). I found it strange that I could queue so much mouse movement and then cause a WM_Destroy (programmed for escape key) which happens instantly but I'm guessing that is given priority. My two initial idea to get around this was to increase the rate at which the messages were checked. My current temporary work around is to use 'D3DPRESENT_INTERVAL_IMMEDIATE' which puts me up to a very high framerate but gives alot of tearing and won't hold when I start rendering more things or running it on a different system and the framerate drops below 200fps. My other idea was to create a 2nd thread. At first I considered making a different thread to check the messages but it would be much simpler to leave the message checking the way it is and make a seperate render thread. The problem is that I didn't want to do this because I then have to start thinking about critcal sections and such.

Thanks Koobazaur
Thought about not rendering until all messages are dispatched.
I looked up peekmessage on msdn to see if their was a way to check if there were more messages and realised that peekmessage returns 0 when there are no messages so I changed my
if peekmessage....
to a
while peekmessage....

so when there are no messages left it returns 0 and breaks out of the while loop and only then it renders.