Regarding the cache issue, I'm actually curious enough now to look something up: I went through a number of 15" G4 Powerbooks (issued at work) over the years, starting with an original 400Mhz model and ending with that last high-res DDR2 1.67Ghz model. (I seem to recall the sequence was: 400 > 667 > 867mhz > 1.5 > 1.67Ghz.) Honestly the only speed bump where I really recall noticing how much faster the new 'book was than the last one was the 867-to-1.5Ghz swap, which basically confirmed my theory that a computer upgrade should be at least "almost twice as fast" to be worth doing.
Looking at Everymac, the 867Mhz had the 256k+1MB "three level" cache architecture while the 1.5 Ghz was the straight-up 512k on-die. So apparently extra cache won't make up for a 633 Mhz clock speed deficit. ;^) But the difference probably also needs to be put into a historical context: at the time the difference between those two machines was enough that 1.5Ghz was objectively "fast enough" for, say, watching most videos, while the 867mhz was starting to get into "annoying level of frame-drop" territory with newer codecs. (It also *really sucked* at running Java, which I needed for a few things.) That was a real worthwhile/noticeable delta at the time. However, when I last worked with Powerbook G4s a few years ago (prepped a few to give away) frankly I didn't see that once-meaningful difference between a 867mhz and a 1.5Ghz anymore. Both ran 2005-vintage software and play DVDs/mpeg2s/non-H.264 mp4s okay, and both were pretty useless on the Internet once Flash got involved. I guess... trying to extrapolate my experience to what you're seeing between a 1.33ghz iBook and a 1Ghz iMac it seems to me that the list of things that would be "acceptable" on one and not the other today must be pretty short/application specific.
Looking at Everymac, the 867Mhz had the 256k+1MB "three level" cache architecture while the 1.5 Ghz was the straight-up 512k on-die. So apparently extra cache won't make up for a 633 Mhz clock speed deficit. ;^) But the difference probably also needs to be put into a historical context: at the time the difference between those two machines was enough that 1.5Ghz was objectively "fast enough" for, say, watching most videos, while the 867mhz was starting to get into "annoying level of frame-drop" territory with newer codecs. (It also *really sucked* at running Java, which I needed for a few things.) That was a real worthwhile/noticeable delta at the time. However, when I last worked with Powerbook G4s a few years ago (prepped a few to give away) frankly I didn't see that once-meaningful difference between a 867mhz and a 1.5Ghz anymore. Both ran 2005-vintage software and play DVDs/mpeg2s/non-H.264 mp4s okay, and both were pretty useless on the Internet once Flash got involved. I guess... trying to extrapolate my experience to what you're seeing between a 1.33ghz iBook and a 1Ghz iMac it seems to me that the list of things that would be "acceptable" on one and not the other today must be pretty short/application specific.