Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Did Apple take a wrong turn?
- Thread starter Wando64
- Start date
- Sort by reaction score
The 'proper engineering and functionality', as you call it, starts with the 14" MBP. Physics being physics even that's not as well cooled as the 16". But they're both kinda hefty compared to the Air and 13", and for some people, the weight-advantage of the Air and 13" trumps the overall better performance and heat-efficiency of the 14" and 16".
It inevitably comes down to where exactly in the design do you want the compromise, because it has to be somewhere: you can't beat physics.
- Powerful, thin-and-light laptops are not going to be heat-efficient (13" M2 MBP).
- Heat-efficient, powerful laptops are not going to be thin-and-light (14" and 16" MBP).
- Thin-and-light, heat-efficient laptops are not going to be powerful (M1 MBA, M2 MBA tbc).
Last edited:
Checkout the notebookcheck review I linked above. When you hit both he CPU and GPU at the same time, the fans run at max speed and the computer throttles down.
The internal temperature reports are all over the place - likely istat/TGPro/etc need a software update to support the new device - but the behavior is the same in multiple reviews. The device runs at maximum fan speed and begins to through when you push it hard.
The M1 and M1 Pro did not have any problems with throttling (outside of the fanless M1 MacBook Air of course). The M1 Max did have some problems with throttling when you hit both the CPU and GPU (even outside of the 14" MacBook Pro), but that seemed to be a power constraint, rather than thermal constraints.
To add to the disappointing M2 and disappointing chassis, they've lumbered half the world with price hikes. A sad time to be an Apple fan.
I'll be keeping my M1 Max (which took 3 months to arrive!), for as long there isn't a bigger leap in performance or architecture. Such as from Intel to Apple Silicon, even when M2 Pro/Max arrives.
I still think they should have stuck with M1 line for a little longer. The M2 just feels rushed. Yes I appreciate it took 3 years etc, but doesn't stop it from feeling rushed to the end user.....the M1 still seems to be the sweet spot for now....
I still think they should have stuck with M1 line for a little longer. The M2 just feels rushed. Yes I appreciate it took 3 years etc, but doesn't stop it from feeling rushed to the end user.....the M1 still seems to be the sweet spot for now....
Nah, it's the A14 vs the A15 all over again. The A15 has an extra GPU core and they clocked the chip higher - it throttled more than the A14, but even when it's fully throttled, it performs better than the A14.
M2 has more GPU cores and is clocked higher than M1, but the performance is still better. We don't have a bunch of angry people out there with iPhone 13 Pros overheating do we?
M1 and M1 Pro have the problem too. As I have read in a forum the 14 Pro throttles just like the 13 Pro M1 when you put the CPU and GPU under full load. (Cinebench+3D Mark)
This scenario almost never tested.
The iPad Pro will also get M2. It might be clocked lower and thermal thottled by APPLE to keep things cool.
I have a i5/8/512 Early 2020 Air and it did suffer from this. I bought it refurbed from Best Buy at a great deal when they first launched and everyone was returning them because of the heat/throttling freakout. There was a thread on here about cooling mods that was quite cool to see about some of the solutions these guys tried to upgrade the cooling, I myself just went for a thermal pads on the heatsink and under 10.15 and 11.x I saw about 10-20 degrees cooler temps and it GB 5's about 3200. For me it was worth it adding the pad because I did see real world improvements in heat, I do think it heat ups the battery more as my battery health seems to have degraded more quickly. Still a good machine. I have a M1 Air that I use way more now but I still carry that i5 Air with me everywhere I go.
Okay, I go first and make a guess. Tjmax = 108C 🤣
As an aside, some people in the audience seem to think ramping up the fan sooner will help the heat situation in 13-inch M2 MBP. That's also just fantasy. With such as move/change in the fan curve, its effect is very short lived. Given long enough of a sustained task, such as the 8K raw export (last ~20min), the SoC will hit 108C eventually somewhere in between the 20min duration.
If the youtuber's observation is correct, then what happens is that: SoC temperature reaches 108C, and stays there "long enough" for Apple's firmware to intervene and ramp up the fan speed. The fan speed eventually reaches 7200 RPM. And its effect on lowering the temperature is futile. The firmware further throttles CPU/GPU clocks. Only after that, the temperature drops to around 80C. By then the firmware sees the thermal opportunity to boost CPU/GPU clocks, and it does go ahead to do so. Sooner after, the temperature reaches 108C again. The cycle repeats itself until the sustained workload is done.
This is a classical demonstration of inadequate design in the cooling system.
I’d want some corroboration from other sources before I take this report as gospel.
What they have released is the M2 13” MBP and it has a fan. We don’t know if the M1 Air overheats. We don’t even know if the M2 MBP overheats other than this one report.
The M1 MBA does not have an overheating problem under normal use. If you push it to the max you get a small ~10% slowdown. That is absolutely a fine trade off to me to not have a fan. We don’t know how the M2 MBA will do but no reason to think that it won’t perform proportionately to the M1 Air.
As to the M2 MBP and this video, I’d like to see some alternate tests done before we go assuming that there is a widespread problem.
Last edited:
The maximum junction temperature will be a fair amount higher than wherever that sensor is reading. Maybe ask someone from TSMC, Apple, or a hardware reverse engineering firm like Tech Insights. Not that their answer would mean anything to this thread.
Nice one but perhaps you're referring to the temperature that processors will automatically shutdown to prevent meltdown. There is such a temperature but I forgot its name.
From recent usages e.g. by AMD & Intel processors, I see they refer to Tjmax as the temperature when frequency start to throttle.
The junction temperature is the temperature at the junction of a transistor, and the maximum junction temperature is the temperature at which the transistor junction can’t operate within its specs (as defined by the manufacturer).
The sensor is nowhere near whichever transistor’s junction. Its readings are the aggregation of a bunch of thermal materials, sources, and sinks. Dave Jones (EEVBlog) has a good intro video for thermal calculations if you want to learn about that.
I would prefer reading to youtube video. So I found [0] and [1] for the sake of background info.
Take one step back. I was trying to say 1) TjMax won't be a fair mount higher than reported by the SoC, and 2) TjMax isn't something secret that requires privilege to access. You could make a reasonably good guess from the SoC's reading.
For AMD & Intel processors, TjMax values are published numbers. E.g. Ryzen 5000 [2] have some SKUs with 95C TjMax, some SKUs with 90C. Intel processors in the past have TjMax around 100C.
With Intel Macs for example, people can push performance to limit, and watch the processor temperature hovering around 100C, and thermal throttling kicks in.
[0] intro to temperature specification
[1] Intel FAQ
[2] Sample spec of AMD processor with TjMax listed
Nowhere near? On-die temp sensors are diodes - literally semiconductor junctions built in the same layer of the chip as transistors. If you locate one right next to the circuit you want to measure the temperature of, there's not much delta since the silicon substrate is a decent thermal conductor.
“Right next to.” Can you point out where the diodes are buried on die? Everywhere? Also, who cares? This thread is completely beyond the pale.
Last edited:
Seems like Apple is pushing things a little bit more than they should've with the M2 without going for 3nm.
I just hope Apple is not reaching a ceiling so quickly. I mean here's hoping better clockspeed and thermal control on the Macbook Air M2, and the M2 Pro/Max unlocking the real potential of the new cores of the M2.
Wish Apple showed us their roadmap. Intel, as bad as they were, does provide somekind of roadmap on their CPUs so we can have an educated buying decision. With Apple, it's all just a guess. No roadmap, no timeline of support.
For example, Apple is now selling an M1 Macbook Air (2 years old). How long will it be supported till? Consumers doesn't have the information.
I just hope Apple is not reaching a ceiling so quickly. I mean here's hoping better clockspeed and thermal control on the Macbook Air M2, and the M2 Pro/Max unlocking the real potential of the new cores of the M2.
Wish Apple showed us their roadmap. Intel, as bad as they were, does provide somekind of roadmap on their CPUs so we can have an educated buying decision. With Apple, it's all just a guess. No roadmap, no timeline of support.
For example, Apple is now selling an M1 Macbook Air (2 years old). How long will it be supported till? Consumers doesn't have the information.
Take a look at the first generation of Zen die as an example. Below is one CCX with four zen cores. All the stars are sensors/monitoring spots. 20 of those are temperature sensors.
Pick one you like to be Tjmax
The Air is a thin, fan-less computer. If you bought one expecting to push it like a MBP with two fans you're going to be disappointed. If you use the Air in the way the design was intended to be used with light to moderate tasks its great.
I'm not sure why people are acting like Apple planned an entire architecture change and future roadmap based on the performance of the first generation M1. Some M# generations will offer a big jump in performance, while others will have more modest gains.
Last edited:
Honestly it would take some extreme circumstances to get those results. In this case an 8K export. People that do this kind of stuff would go to the 14 or 16 inch Pros in the first place.
Sadly MaxTech kinda lost its way. It used to be a great channel. But lately with the daily uploads and their anxiety to get to 1 million subscribers, they tend to clickbait and make a drama more and more.
Sadly MaxTech kinda lost its way. It used to be a great channel. But lately with the daily uploads and their anxiety to get to 1 million subscribers, they tend to clickbait and make a drama more and more.