I learn a lot from Linus Tech Tips. His opinions are often very different than my own but he's incredibly knowledgable and passionate and, whether I ultimately end up agreeing with him or not, the process forces me to evaluate, re-evaluate, and learn. And that's absolutely the case with one of his latest videos: Macs are Slower Than PCs video. And Why.
So, is that true or is that just his point of view?
Now, please keep in mind that I'm a long time MacBook Pro user and someone who generally values the trade-offs Apple makes. So, that's my bias, but I'm going to try and keep this as balanced and fair as I possibly can, and I'll trust you to shout at me in the comments if and when I get anything wrong.
So, the crux of Linus' video is the question: Why Macs are slower than PCs. Now, for the sake of completeness, PCs are more of a category than a product line, and technically Macs are PCs, so it's a little like saying why Ferraris aren't faster than cars. They're faster than Carolas, maybe many not Lambos, but not custom jobs with warp nacelles strapped to the sides. But, for the sake of argument, I'm going to stick to similar laptops and desktops made by similar vendors.
Now, Linus is the benchmark master here, not me, so for a general evaluation, I'm going to go with what he said in his previous video about the current generation MacBook Pro: That the new i9 MacBook Pro is close to or faster than his other i9 reference machines.
Mac-targeted title of the video aside, what Linus is really arguing here, or at least one of the major arguments he's making, is that despite the prices being pretty much the same, spec for spec, between Mac and PC, Macs often don't perform as you would expect given their specifications.
He's all hubbah hubbuh over the designs and specs but feels, and this is a totally familiar refrain, that the designs — or in his words the sex appeal – don't do justice to those sometimes very impressive specs — because thermals.
Linus shows some stress tests and says Apple's done a little bit of trickery, adjusting the voltage and fan curves so that they can hit a higher thermal threshold without throttling too far below Intel's advertised base clock.
I'll quibble with this part a little. What he calls trickery I think would far better be called management. Given the form factors involved, managing processor speeds to manage temperature just makes the kind of sense that does.
Linus points out that this management disappears under Bootcamp for Windows, and there was an early bug that prevented it from working and caused a ton of controversy on the Mac as well, but that was fixed pretty damn quick. Still, Linus asks, again, if it's all just software trickery to mask a bad design.
But here's the thing: It's part of the design. Apple's whole thing, since the beginning, has been about software, hardware, capability, and design all working together. All design is compromise, all engineering is trade-offs. A giant, thermally unmanaged battleship you can't easily or enjoyably carry around or use unplugged is also compromised. Just differently compromised.
The high order bit, and they said it on Reddit, is this: sustained speed, reduced noise, or realistic portability, pick two. But the next highest, the one that people don't talk about as much is that, in some cases, you can manage, mitigate, and even engineer the hard edges off your compromise.
Now, Linus says PC laptop manufacturers thermal throttle as well, though he says many, which I take to means not all, to a lesser degree. That makes it sound more like a Why Are PCs Slow Video, rather than anything Mac-exclusive.
Linus also points out that all of this has been exacerbated by Intel's multi-year failure to shrink their process down too 10nm, and their compensating by, once again, throwing more cores at the problem, resulting in higher temperatures that manufacturers have to try and over-prepare for.
Which, again, can totally be done, but not without other, different compromises on the aforementioned noise or portability or dependency on AC rather than battery power.
Some people might prefer those other compromises, others people decidedly not. Because here's the thing: Not all customers are the same. We all have different needs, certainly, but also different priorities and even different tastes.
Now, on the internet, we have a special name for people who have different preferences than our own — idiots.
Yeah, it's funny because it's true. If you like Macs and iPhones you just don't understand how people can deal with those ugly, incoherent Android or Windows boxes. And if you like Android or Windows, you just don't get how anyone can put up with those locked down, design-constrained iPhone and Mac boxes.
It's not like we're all speaking different languages, it's like all our brains are all wired differently. And, they are. On the internet, we just get super tribal and shorthand that down to everyone else being, you know, idiots.
Linus explains just as much when he says that, for a large enough proportion of Apple's customers, the looks and the status symbol of owning the machine are just more important than whether it's actually quick off the line.
And I get it. I totally get it and for exactly the reasons I just mentioned. Apple isn't solving for just one attribute for just one percent of the customer space. They're trying to solve for multiple attributes, to balance power with portability and quietness, in a durable and, yeah, well-designed package.
Apple is managing for thermal, sure, like other vendors. More in most cases. They've been doing it for years and so far, there's no real-world evidence to suggest it's been a problem. Not at all. But Apple's also been engineering around it by including top-of-the-line SSD and custom controllers. They're even including custom silicon like the T2 chip in more and more models, which handles security and device control, which helps with performance in several ways, from encryption speed to encode and decode.
Because, and you know it's true because it's already a cliche, it's not the specific number of the benchmark that really matters when you own the machine, it's the overall speed you experience every day while using it.
There are certainly some customers who prioritize aesthetics and want people to see Apple computers on their desks, in their bags, at their reception areas, and throughout their offices, the same way they want specific chairs and paintings and lights and whatever. But, even then, it's not like that brand value just magically appeared.
Apple's nailed that part. Hard. So hard, an entire generation of other notebooks have raced to follow Apple's designs. And that's not a knock. That's credit where it's due.
For others, it's the materials, like aluminum and glass, and their durability over time. Point being, Everyone has their own idea of overall value.
Linus goes on to say that the advent of turbo boost has quote unquote masked Apple's negligence. He says turbo boost allows Macs to feel pretty snappy in day to day use, But, he also says, they require literal water-cooling to reach peak performance in heavy workloads.
But, again, I'll quibble with his characterization. The vast majority of people don't want or need a literally water-cooled MacBook Pro. They want and need exactly the snappy feeling one in day to day use Linus mentions.
For them, Turbo boost works. It's a great way to address a wide swath of bursty workflows that benefit from exactly that kind of performance. Things like opening apps, which, in the age of SSD, is no longer IO bound. Browsing the web. Making adjustments in Lightroom, trying out filters in Photoshop, anything quick and intermittent. That might not sound important but you add all those bursts up and they equal real time, especially for pros for whom time is worth far more than money.
That's turbo boost doing exactly what it's supposed to do.
Linus also kinda asks why Apple considers it ok for their pro Macs to throttle in a way the iPhone doesn't. Why this difference in philosophy?
Again, he knows much more about this stuff than I do, but I'll do my best to fill the cricket sound with a pretty good guess: The consideration and philosophy are the same: to deliver what they think is the best product for the most customers. The only real difference is that Apple custom matches the silicon for that product on iPhone and iPad. And they don't, at least not yet, for the Mac.
Linus thinks Apple is only throwing the biggest chips in the smallest packages for what he calls, Le Marketing.
Jonathan Morrison did a pretty comprehensive set of comparisons between the 2018 i7 and i9 MacBook Pro, so I'll just give you his conclusion: The i9 is still faster.
Perhaps even more telling though is what Apple did with the 2018 MacBook Air: Launched in with only one processor option. And, of course, some people were upset by that decision too.
But, rather than Le Marketing, it could also, possibly, just maybe, be that Apple is trying to do what it thinks is best for each product line, including working around Intel's lack of process shrink as much as they can — and as we learned this week, constraints on even the chips Intel is able to ship out — including at the highest end possible. Because, as we saw back in 2016, if Apple doesn't offer those highest end options, they get castigated for that as well.
So, why is the Mac slow than the PC?
No, of course not. Despite what Linus says, what he shows, over and over again, is that it isn't. He's wrong about almost everything in the video and the one thing he's right about, that the thermals prevent maximum performance, applies to similarly designed PCs as well, which makes it decidedly not a Mac thing at all.
Now, yes, if we want to artificially constrain the conversation to maximum sustained performance, as though clock speed exists in a vacuum, that the speed benefits of the other components, of macOS and Final Cut Pro X, of everything else doesn't matter, then sure: Apple could make different design compromises to prioritize that. But it wouldn't be for free. It would cost those other compromises, including size, weight, and fan noise, compromises that might not appeal to the vast majority of Apple's customers.
No product is the sum total of one dimension or spec, and when we forget that, or assume our own preferences and use-cases should be everybody's, even most people's, and when we can't perspective take or empathize with the needs of real people in the real world, that's when both customers and audience lose out.