iTunes Speed Feedback

[William Hofius]

Just a couple of additional comments:

1.) While I can understand your disappointment with the iTunes 4 performance, I do not think it is appropriate for you to complain to Apple about feeling "ripped-off" for a decision you made regarding which processor to buy. You can yell at Apple for feeling ripped off with iTunes though! :-)

2.) Based on the information I sent in my previous message, I would say that any use of dual processors that iTunes 4 via ACC is making is based on OS X system resource allocation and not iTunes 4 or the ACC encoder. When ripping MP3s on my dual processor Mac, each processor is utilized equally and consistently. (I have not tried ACC yet myself.)

3.) When I got my 800 MHz TiBook, I was expecting a nearly doubling of performance was well. I run the command line version of SETI at Home. On my 450 MHz DP PowerMac, I run one instance of SAH per processors. It takes my PowerMac about 12.5 hours to complete one work unit. I was expecting my 800 MHz TiBook to complete an SAH work unit in about 7-8 hours. It takes my TiBook about 10 hours to complete an SAH. I was a bit disappointed. I did some research and this is what I found out.

     •While L3 cache is nice, L2 cache is what's important. When it comes to L2 cache, while speed is nice, size is what is important. While Apple says that 256kb of L2 cache at full processor speed performs as well as 1 MB cache at 1/2 or 1/3 processor speed, for big number crunching this ends up not being true; a larger cache. The old 500 MHz CPU most likely had 1 MB of L2 cache. The new CPU upgrade probably only has 256kb of L2 cache.

     •Cycle for cycle, the new generation of G4 processors are not as effcient as the original G4 processors. Remember the Motorola G4 processor snafu? G4 chips could not be clocked faster than 500 MHz. (Well they could, but Motorola could not produce enough error-free chips.) In order to get around that issue, Motorola redesigned the G4 architeture in order to increase the chip's clock speed. (I think it went from a 4 stage pipeline to a 7 stage pipeline.) While this enabled faster and faster G4 clock speeds, there was a trade-off in processor efficency. Cycle for cycle, your old 500 MHz gets more work done than your upgrade.

Ultimately, the fault is not iTunes 4.