We couldn’t care which chips are in our computers. As long as they do the job they could be any brand. After all, we’ve lived with two or three Zilogs, a MOS Tech, umpteen Motorolas, and two matched pairs of IBMs. Each was better than their predecessors and when new, seemed amazingly more capable.
By a bit of inspired forward planning, we’d got dual-processor computers before the operating system would support them. This was a year prior to Apple’s switch to a Unix-based OS and when it did, our Macs just flew. They could run just about every application we had, all at the same time. This was even though half were still running in the emulated Mac Classic – which ran them with more stability than native Mac OS 9. As far as we were concerned, any chip was as good as the others. Then we got our first Intel computer.
Intel-igence
When it arrived our MacBook seemed to be rapid and capable, if a little hot. Give it a task to complete and it zooms away, the Core2Duos seemingly faster than a same-speed G5 twin processor. Until, that is, we piled on a few more tasks. Then the computer becomes unresponsive at times, the pointer doesn’t move and you get the feeling it couldn’t be bothered to put up a spinning beachball even if it had time to do it.
Testing time
We decided to delve a little deeper and compare the Macs we had on hand to see whether our feeling the Intel CPUs were bogging down was correct when compared with our Motorola and IBM powered alternatives. For this we looked at the total maximum tasks each is capable of plus the number available to the user. These are set at the time applications and systems are compiled although they can be changed temporarily. It should give a rough idea how the CPUs will handle multi-tasking, dealing with the frontmost while at the same time getting on with all the System applications and any other jobs it has been given to do.
The testing was done simply by entering sysctl kern.maxproc in Terminal to find out the maximum tasks the computer can handle and sysctl kern.maxprocperuid the maximum available to the user. Enter man sysctl kern in Terminal for more details.
A big surprise
Our results were surprising. A G5 2.3 GHz dual processor gave exactly the same result as a G4 800MHz single processor, each is capable of 532 total maximum processes with 100 available for the user. The Intel 2GHz Core2Duo achieved 532 maximum with 266 for the user so it should be as capable if not doubly so. Finally, our first generation G5 2GHz twin processor is the biggest surprise of all. It can handle 512 maximum processes with a massive 512 available for the user.
This reflects our real-life experience with the computers. The little G4 iBook is an excellent laptop for general tasks and despite its relatively slow chip (by modern standards) is capable of just about anything thrown at it. The G5 2.3GHz flies through heavy-duty Photoshopping with one hand behind its back while the G5 2GHz is our best server and all-round workhorse. It can compress movies to H264, at the same time streaming live TV and playing the same shows, download mail, web browse, file serve, housekeep, back-up, virus scan and all the rest we throw at it, yet it still plays a mean high-res Call of Duty. All at the same time. It’s fans spin a lot more than the others but not intrusively so.
Heard it on the pipeline
All the blurb about pipelines, threads and other geeky jargon was just that as far as we were concerned. For all we know, all the above may be a load of rowlocks. But it seems to us that Intel’s CPU’s are brilliant at doing single tasks very fast but not so good at doing loads of things all at the same time. Didn’t Apple have a video of Avi Tevanian showing blue blobs moving through tubes and demonstrating shorter tubes move blue blobs more quickly than longer ones? He explained this was why the G5 is better than Intel chips – Big Blue moves blobs more easily.
That all seems to have been forgotten in the hype about Intel inside Macs.