One of precious few pieces of good news last week was Apple’s MacBook Neo.

On the face of it, the Neo is the Apple silicon Mac with worse specifications than the first base M1 models of 2020, with a mere 2 Performance and 4 Efficiency CPU cores and a 5-core GPU, just 8 GB memory and 256 or 512 GB internal SSD. And it doesn’t have a single Thunderbolt port. But if you’re buying your Mac on its technical specifications, you’re not going to be a Neo user.

For consumers, the education sector, and the many others who want something lighter than Air, and who don’t want to pay for all the features they’ll never use, a Neo will be ideal. The only shortcoming I can see isn’t in the Neo itself, but in backing it up. As Time Machine doesn’t back up to iCloud, wouldn’t it be ideal if Apple were to offer a wireless backup system? Until eight years ago, it had what was then a perfect product in its Time Capsule.

If the Neo is as successful as it deserves to be, it’s likely to revitalise many independent software developers, who offer neat little apps rather than the heavyweights rented to us by large corporations, including Apple Creator Studio. Maybe we’ll even see the return of a compact office suite like AppleWorks.

I’m more cautious about the announcement of a third CPU core type in M5 chips, largely because of the current lack of detail. When it comes to evaluating the high end performance of M5 Pro and Max chips, the devil really is in those details.

For the M5 family, Apple has apparently switched to three types of CPU core, instead of the Performance and Efficiency types that have proved so successful in the M1 to M4 families. Now we have Super (S), Performance (P) and Efficiency (E) to juggle with instead.

Apple claims M5 S cores are the “world’s fastest CPU core for single-threaded performance”, with their increased front-end bandwidth, new cache hierarchy and enhanced branch prediction. From previous measurements, they’re expected to operate at frequencies ranging between 1,308-4,608 MHz, as the four in the M5 base chip do.

Next are regular P cores, claimed to be optimised for power-efficient multi-threaded workloads, and E cores for running all those background threads whose economy is more important than speed. Those E cores operate at frequencies between 972-3,048 MHz, and they have about half the processing capacity of regular P cores.

Ignoring binned versions, the M5 family now has three members:

• M5 base 4S + 6E
• M5 Pro 6S + 12P
• M5 Max 6S + 12P

The biggest difference in processing between the Pro and Max are their GPUs: M5 Pro chips come with 16-20 GPU cores, while Max chips double those to 32-40. For a price difference of around $/€/£800, those GPU cores seem expensive. If you’re considering either chip, be sure to price up equivalent systems using each, and ask yourself whether the additional cost of the Max is worth it. I suspect that, as with its M1, Apple hasn’t put sufficient distance between the M5 Pro and Max for prospective purchasers.

What we don’t know yet is how macOS manages the frequencies of M5 S and P cores. One commonplace situation that merits close examination is the initial phase of 5-10 minutes background activities following user login. With a choice between P and E cores, macOS runs most of those, including Spotlight index maintenance, on the E cores. In their absence, those will have to run on P cores instead, where they may contend with user interactive threads.

Existing apps guide macOS in its choice of core type in which to run threads, using a Quality of Service (QoS) value assigned by the developer. macOS will then try to run threads with higher QoS on P cores, and those assigned a low QoS are normally constrained to run on E cores. That can work well in previous Pro and Max chips, where there are many more P than E cores, but a different interpretation of QoS is going to be necessary for the M5, where S cores are the limiting resource, and apparently best-suited to running single threads. Apple hasn’t yet released any information or guidance to developers.

This will require more detailed studies than merely comparing the usual benchmarks, and I suspect there may be some situations where the new M5 architecture won’t be as clearly beneficial.

Between the MacBook Neo and M5 Pro and Max chips, we have exciting times ahead.