Over the last week, the press coverage of Apple’s new Mac mini, MacBook Pro, and MacBook Air — all powered by the company’s new M1 SoC — has been glowing. ExtremeTech’s coverage has focused on the M1 as a CPU and its potential for market disruption rather than diving into the day-to-day experience of using the new Mac systems. We’re plenty excited about what the M1 could mean for personal computing and the products from other manufacturers it might inspire, but the practical side of the equation matters to anyone who might, you know, spend money on one of these things. To that end, it’s worth noting that Apple’s ARM-macOS ecosystem is still very new and perhaps a bit less polished than Apple’s PR makes it sound.
According to Patrick Moorhead of Moor Insights & Strategy, there are enough warts and rough patches to the entire affair to make caution a well-advised strategy for now. He writes that the M1 “will be fine for users who use 100% Apple software, stay primarily in Safari and don’t need to connect it to a bunch of peripherals, albeit pricey.” For other users, it’s a bit more complicated.
Moorhead lists a number of applications that have given him trouble in recent days, including Edge, Outlook, and Logitech Camera Control. Winzip 8.0 refused to open after installing, and backup software from Samsung malfunctioned. Adobe Reader XI 11.0.10 refused to install. It turns out there are some curated lists of supported software and a fair number of applications that are not yet supported. Even when products are supported, they may have various unspecified issues — a number of Adobe products fall into this category at the moment. Avid Pro Tools is another application that doesn’t support the M1 yet. Neither does Pixelmator Pro, though that application is supposed to be updated this week. HP printer drivers also failed to install properly.
These issues sound similar to the problems Microsoft has/had with Windows on ARM when emulating software and attempting to attach third-party peripherals. Apple will obviously try to squash these sorts of problems as quickly as possible, but peripheral support could come down to the willingness of third-party manufacturers to write new drivers.
Another important point Moorhead raises concerns battery life. While the laptop runs beautifully in native apps, it may draw far more battery power in emulation mode. Moorhead reports a battery life of 4.5 hours while running “Outlook, OneNote, Chrome WhatsApp, Word, and performed one Skype for Business call, one Zoom call, and one Webex call.” This is about half of Apple’s claimed battery life.
Battery life over time. Image by Pat Moorhead.
It would not surprise me if follow-up tests validate this figure. Emulation always has an overhead, and Apple may have chosen to prioritize performance over power when in x86 emulation mode. The 10,000-foot reason for this is because the x86 and ARM instruction sets do not correspond to each other in perfect 1:1 fashion. While the details differ depending on the type of emulation (hardware, software, firmware, etc), there’s an inevitable performance and power penalty when translating from one instruction set to another. Apple could theoretically address this by running the M1 at a higher clock rate when emulating as opposed to when running native aps, or it’s possible that the specific workload Rosetta 2 creates on the CPU requires it to perform additional calculations that burn power without affecting clock. Either way, the idea that emulation could hit battery life hard is not surprising. The battery life hit will be application-specific.
Moorhead raises some additional points regarding the overall value proposition of the ARM-based MacBook Pro and we think it’s worth a read whether you’re over the moon for these specific systems or not. He splits the difference himself, coming down as impressed with the M1 but less happy with Apple’s messaging relative to his actual experience with the product as a daily driver.
We’ve said several times that we don’t expect the M1 or ARM more generally to devour x86’s market share overnight, even if the chip proved to be extraordinary. This type of low-level software and battery life cruft is one of the technical reasons why. There are going to be applications and scenarios where the M1 will either fall short due to slower emulated performance or will draw far more power than an equivalent x86 chip to complete the same task.
We don’t think the x86 manufacturers can pretend those issues are going to be a long-term problem, but it’ll probably be 12-24 months before the entire ecosystem switches over, and there will inevitably be some companies that drag their feet. It’s clear that some users will get more out of transitioning to an M1 system than others, and it’s always a good idea to make sure your personal usage model is a good fit for the OS and available software before taking the leap on a new piece of hardware. In this case, it seems the software and battery life situation with the new MacBook Pro may be a bit more nuanced than previously reported. That’s not unexpected, but it’s not a message that came through very clearly in early coverage of the three specific systems, either.
Now Read:
- Why Apple’s M1 Chip Could be a Real Threat to Intel and AMD
- The New Apple M1 Reviews Put AMD, Intel Officially on Notice
- Apple’s New M1 SoC Looks Great, Is Not Faster Than 98 Percent of PC Laptops
No comments:
Post a Comment