So who's tested CS3 vs. CS4?

Question

I haven't seen any tests yet, like the standard rotate blur test from retouch artists
http://forums.macrumors.com/showthread.php?t=200558
or the more involved one at driverheaven:
http://www.driverheaven.net/photoshop.php
Anybody care to post results of the two versions using the same settings and hardware? I'm very curious how much faster CS4 is, especially on a MacPro.

Answer 1

Neil,
I agree that those simple tests don't prove much. The test that Adam referred to on MacGurus supposedly gives a better assessment of speed so one can tweak hardware.
I do think a standardized test is helpful in determining how fast a version of PS will run on your system. The longer of the two tests, Retouch Artists action one, should give some decent info, but I think for users the best way is to evaluate the software using actions they use on batching lots of images and/or how they retouch. My actions included bit-depth changes and color space changes as well as saves, and I'm suspecting that adobe has improved the save speed. Whatever they've done in CS4, it is appreciably faster than CS3, making my 2.66 quad feel more like a 2.8 quad, which makes me happy indeed.
PECourtejoie,
You are right about the 1 history state versus 20 history states. 1 history state does tell you something about the speed of the hardware, but it is not what regular users are using. These tests only give you a limited window into real world performance. At best, they are used for bragging about hardware. Users who are actually interested in real-world productivity need to test with actions that use all aspects of photoshop that they may use in their work. As my testing proved, I did not see any improvement with Retouch Pro or Retouch Artists in CS3 vs CS4, but when I ran batches using identical actions on the actual files that I use in my work, I discovered that CS4 actually was about 20% faster.
Paul