It wouldn’t be an Apple product launch without some doom-mongering, would it?
This week, Max Tech’s Vadim Yuryev used lots of uppercase words to reveal his latest benchmark test results for the M2 MacBook Pro.
It’s bad news, I’m afraid, folks. That thing throttles, big time, and far more so than the laptop of the same name with the M1 chip. As Yuryev demonstrates, this is particularly the case when you’re exporting 8K RAW Canon footage.
Yeah. I know.
This is starting to get a bit silly, isn’t it?
This is all Apple’s fault
I’m not about to pull Yuryev’s findings apart, nor am I going to criticise what he does. Like me, he’s a content creator vying for attention; we just go about it in very different ways and have entirely different strategies for our content. Neither of us is doing anything wrong. We just find different elements of tech interesting.
Instead, I would like to point the finger of blame for this rising tide of bad press for the M2 squarely at Apple – but not for the reason you might expect.
When the M1 chip launched in 2020, Tim and co. began beating their chests so hard you could hear the hollow echoes emanating from behind their rib cages on other continents. Then came the Pro, Max, and Ultra variants of that chip, and the chest-beating grew so loud that it began to escape the Earth’s upper atmosphere.
This gave the M2 chip a huge mountain to climb. It had to somehow continue this trend of wiping the floor with the PC competition and reinventing computing as we know it.
That was never going to happen. Whether or not the M2 is demonstrating the current limits of silicon which is, to all intents and purposes, designed for smartphones, is irrelevant. It’s nothing more than a marginal update to the M1; of course it’s going to struggle under load.
The issue is these damn benchmark reports. This isn’t Max Tech’s fault – it’s Apple, once again, baiting those benchmark obsessives into diving as deeply as possible into every aspect of that new chip.
“Oh, so you think you’re ‘all that’, eh, Mr Apple? Well let’s find out if that’s really the case, shall we?”
I totally understand why the likes of Max Tech do this. But it’s exhausting. It’s boring. And I think it’s a huge problem for the vast majority of potential MacBook owners.
The test results
These aren’t the first less than satisfactory test results for the M2 MacBook Pro. As reported recently, it appears that the base model of that machine (the one with 256GB of storage) has slower read and write speeds than its predecessor.
This time, however, the testing was focused entirely on the raw performance of the M2 chip. And it doesn’t look great.
“We discovered SEVERE thermal throttling with Apple’s new M2 MacBook Pro, proving that it needs a BETTER cooling system with two fans instead of one,” explained Yuryev.
He was referring to Max Tech’s Final Cut Pro export test results for the M2 MacBook Pro (the exact model they used is, unfortunately, unclear). While exporting 8K Canon RAW footage (yeah, I know, I’ll get to that in a moment) the M2 apparently hit 108°C. Worse still, it spent the entire export with its single cooling fan running at full speed, thus leaving it with no choice but to throttle the performance to keep the temperature under control.
As Yuryev points out, this had a significant impact on the power output of the M2 during the render process.
“In a split second, the M2 chip would cut its P-core clock speed from 3200MHz to 1894MHz, its E-core from 2228MHz to 1444Mhz, its GPU from 1393MHz to 289MHz,” explains Yuryev. “This resulted in total package power dropping from 29.46W to 7.31W.”
I have no idea what any of that means, but it certainly sounds like the M2 chip in his M2 MacBook Pro was having a rough old time.
These dips in power output would apparently happen in “WAVES”, with the M2 attempting to rapidly cool itself before stamping on the throttle again to get back up to full speed. The result? An export time that was over eight minutes longer than the M1 MacBook Pro.
The moral of the story as far as Yuryev is concerned?
“It’s not looking so good for the M2 MacBook Air.”
I disagree. And I think, once again, we all need to chill out a bit.
Why a HUGE pinch of salt is required
This blog post isn’t an attempt to defend Apple. Regular readers will know that I’m not averse to calling them out if they do something stupid (which happens quite a lot).
But this M2 stuff doesn’t fit within that category. These early benchmark test results for the M2 chip aren’t indicative of Apple losing its mind, pulling the wool over the eyes of consumers, or failing miserably. It’s nothing more than the continued evolution of Apple silicon, and it ain’t always going to be plain sailing if you know what to look for.
I understand Yuryev’s concerns about the M2 MacBook Air. That laptop doesn’t have a fan, and it is therefore entirely reasonable to draw the conclusion that it’ll consequently throttle the performance of the M2 chip even further.
However. Let’s go back to the subject of that benchmark result, shall we?
Max Tech chose 8K RAW Canon footage as the benchmark for their tests. Although I have no evidence with which to back this up, I can confidently say that no one will use either the M2 MacBook Pro or M2 MacBook Air to work with that kind of content. Anyone working within that level of production will be using a Mac Studio or a Mac Pro. Simple.
If, on the other hand, this test had been carried out with regular 4K footage, we’d have something interesting to hang our hat on. I think owners of these laptops will use them for that kind of work, even if they’re in the minority. But Yuryev and co. probably know that a test of that kind wouldn’t yield headline-grabbing results. As a content creator myself, I get that – totally.
However, even if 4K benchmark tests did yield similar results, it still wouldn’t matter. Buyers of the M2 MacBook Pro or M2 MacBook Air are highly unlikely to be working against time constraints with rendering and exporting performance for video production. Once again, if they were, they wouldn’t be buying these laptops.
Yuryev proves this later in his tweet thread.
“Just for reference, the base 14″ M1 Pro MacBook Pro didn’t throttle like the M2 did, with the CPU and GPU running at full blast the entire time,” he revealed. “That’s because the fans were able to keep temps low enough to not throttle. This is how PROPER cooling behavior looks like!”
Yes! It is, Yuryev. I agree! And it’s also what that particular MacBook is designed to do. These initial M2 MacBooks aren’t intended for this kind of intensive, sustained, high production value work; they’re built for solid single-core performance and ultra power efficiency.
I’m afraid this renders these test results entirely irrelevant for anyone but those who enjoy devouring benchmark results.
Does this spell disaster for the M2 MacBook Air?
I must repeat: I have no issue with the line of work Yuryev and the team at Max Tech have chosen. They’re very good at it, and, usually, it serves a great purpose.
Sometimes, however, it does more harm than good.
If these benchmark results and Yurev’s hyperbolic tweet threads make it into the world of mass media and consumer news outlets, it might prevent potential M2 MacBook Air buyers from making a very sensible purchasing decision.
“Are you still going to buy that M2 MacBook Air, Dave? Haven’t you heard about the test results that show it’s loads slower than the M1 version? They reckon it’s going to overheat and melt your desk, too!”
If you’re looking forward to buying an M2 MacBook Air, please don’t let these benchmark results put you off. It will be a superb laptop.
If you’re editing 8K RAW Canon footage… then, well, you’re probably not reading this blog post. So, I’ll shut up.
[…] won’t go into the full details of the controversy here, because I’ve written about it already. But, to summarise, a well-known YouTube channel that focuses its content on detailed benchmark […]
[…] Apple’s M2 chip. This blog isn’t the place for my thoughts on that topic, but suffice to say I had quite a strong opinion and wanted to let the world know about […]
This is like saying people who buy mobile with SD 8 Gen 1 shouldn’t complain about the throttling, as their task is so heavy that it should be done on a laptop.