I've been running some various benchmarks from various images and have
the following results using the new 3.5.1 windows vm:

image benchmark benchFib tiny
3.6   151515000 3603990  154124021 bytecodes/sec; 3524771 sends/sec
3.5   156250000 3571227  154309825 bytecodes/sec; 3565153 sends/sec
3.4   156250000 3571227  152562574 bytecodes/sec; 3502103 sends/sec

As you can see they are all pretty close.  These numbers were retrieved 
after multiple runs, by the way.

However, FrameRateMorph reports very different results:

image FrameRateMorph
3.6   20 mSecs (50 frames/sec)
3.5   20 mSecs (50 frames/sec)
3.4   2 mSecs (500 frames/sec)

As you can see, vastly different.

All of the images are set to 32 bit big endian.  It's not the vm, 
because I received the same numbers on the April 2003 windows vm as well.

The FrameRateMorph code seems to be the same across all of these images.

The system "feels" about the same speed and responsiveness as well, so 
I'm wondering if FrameRateMorph is picking up on something different 
between the images, rather than their actually being a true difference 
in speed.

Any help or advice.

Thanks!

Glenn Ehrlich