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.
Cinebench causes my 2018 2.6ghz i7 to hit 100c consistently when running
- Thread starter johnalan
- Start date
- Sort by reaction score
No it’s not normal to run Cinebench on a thin and light laptop. People do that on forums and clickbait sites because reasons.
Well, it serves a purpose to identify issues. Without published Cinebench results Apple wouldn't lift a finger to prepare a throttling fix. if his 2.6 hits thermal limit at 3GHz I'd say it is not normal, should be 3.2 - 3.3 resulting in 1050+ score. If the '3Ghz or so' is actually 3.2 or more then it is normal and expected.
The CPU is advertised and guaranteed to run at the base frequency for prolonged periods of time. So if he gets more than 2.6Ghz, that's normal. Anything beyond that is a more or less random free bonus.
Yes it is good to point out the issue but it would have been fixed. It’s not a practical test to use workstation rendering apps on a thin and light anyway. Especially an old one with deprecated and inefficient APIs like OpenGL. We expect close to metal APIs like Metal or Vulkan to perform better with lower power consumption.
I think to OP's point, I don't know how long you were running the Cinebench results, but it is entirely possible that that is normal temps. As others have pointed out Cinebench isn't really designed for use on a very slim laptop, other than for maybe a quick test once when you first got the machine.
Of course its normal temps. Its the maximal allowed operating temperature for the machine. No wonder that you reach it if you try to push it as hard as it goes.
This has nothing to do with Cinebench. Any kind of CPU-intensive workload on a MBP will bring the CPU close to 100C.
You mean sustained workload such as rendering, but that's no guarantee either. Rendering engine, API efficiency, ambient temperatures and workload content are factors too.
Interesting thanks. I expect though that the MacBook Pro is well able to handle heavy loads. Yes this test was synthetic, but it's a 'Pro' machine. Is consensus here that you shouldn't run it hard because its 'thin and light'?
No, you can use the machine as you see fit, just keep in mind that because it is a small and thin laptop, it lacks sufficient cooling that a much larger desktop with the same class of chips would have and this will cause the computer to heat up considerably at times. Whether or not you think it’s good to run the machine that hard is up to you.
MacBook Pro is perfectly able to handle heavy loads. And you can use it in any way that suits your needs. Of course, if most of your usage is running sustained heavy loads, then I'd ask myself whether a laptop was a good choice to begin with.
Again, don't get confused because of the high temps. 100 degrees is a (maximal) safe operating temperature according to the CPU's manufacturer. If you want more details: these CPUs are assured by the manufacturer to draw 45 Watts under sustained load while running at least at their base frequency. So these are the numbers around which Apple has designed their cooling and power system. It's easy enough to confirm using Intel Power Gadget that the CPU will max out around using 45-50Watt in sustained operation — and it will also reach its maximal allowed operating temperature at the same time. All in all, it's a careful balancing act — designing the smallest possible cooling system that will allow the CPU to operate exactly within its declared spec.
And sure, a more powerful cooling system + removing the 45Watt restriction would allow the CPU to use more power in sustained operation and thus have better performance. Thats how desktop computers operate, and there are also some large gaming laptops capable of this. Apple prioritises the compact form factor while delivering very high performance. This, again, is all about balancing and tradeoff. For instance, MSI GT75 Titan delivers 20%-30% better CPU performance in Cinebench (using the i9 CPU) than the MacBook Pro, but the MSI is also 2.5 times heavier (at 4.6 kg) and over 6 times larger (looking at total volume). In fact, its power brick alone weights almost as much as the 15" MBP! So there are diminishing returns
No, I mean any CPU-intensive workloads. Which is formally defined as running multiple threads that execute non-NOP instructions (that ideally load most of the CPU ports) across all the existing cores, without sleeping. Rendering could qualify as such operation — assuming, as you said, the rendering engine is able to maintain high CPU utilisation. But it could be just as trivial stuff as adding 1 and 1 and discarding the result. Ambient temperatures are of almost no consequence. If you have threads that continuously dispatch work, the system will try to boost the CPU as hard as it can, which will ultimately end up in it heating its thermal ceiling on a system like MBP which has one (it would be different on a system that is designed to cool more then the CPU can ever output — in that case the CPU will run cooler, but also such cooling system would be rather inefficient).