Let's check throttling of 2017 MacBook 12"

[Collo]

I noticed a lot of people want to know if there are any throttling issues with the 2017 12inch MacBooks.

Shall examine this ourselves by performing the same "22 runs of the Cinebench Multi R15 test" done by Notebookcheck.net last year.

That test famously (or erroneously?) put m5 above m7 performance.

We'd need people with m3, i5 and i7 2017 MacBooks to perform the same test and we're golden!

Who's up for the challenge? I'll post mine when my Macbook arrives too!

 

[MagicBoy]

So not the final word then.

I hate clickbait threads.

 

[Collo]

So not the final word then.

I hate clickbait threads.

Better now?

 

[Koudspeel]

So how does this test work? What do I have to install and what do I click and where do I upload to? I've got the i7.

 

[EugW]

Better title now. Throttling review already done by Notebookcheck for 2017 MacBook m3.

https://www.notebookcheck.net/Apple-MacBook-12-2017-Laptop-Review.230656.0.html

Gradually throttles in Cinebench for about 15 minutes, and then crashes through the floor for one run, and then gradually throttles again. However, performance before the crash is on par with the 2016 m5 in your first graph, which is the fastest model in that graph.

 

[Collo]

Better title now. Throttling review already done by Notebookcheck for 2017 MacBook m3.

Then we've got m3 down, i5 and i7 to go!
[doublepost=1500132163][/doublepost]

So how does this test work? What do I have to install and what do I click and where do I upload to? I've got the i7.

Great! Can you grab Cinebench and perform the same test as mentioned here?

https://www.notebookcheck.net/Apple-MacBook-12-2017-Laptop-Review.230656.0.html

 

[konradsa]

Better title now. Throttling review already done by Notebookcheck for 2017 MacBook m3.

https://www.notebookcheck.net/Apple-MacBook-12-2017-Laptop-Review.230656.0.html

View attachment 708725

Gradually throttles in Cinebench for about 15 minutes, and then crashes through the floor for one run, and then gradually throttles again. However, performance before the crash is on par with the 2016 m5 in your first graph, which is the fastest model in that graph.

Yeah, it's pretty amazing that this year's base model is pretty much just as fast at the top model last year (assuming the m5 last year was the top model based on the test above). Got the 2017 m3 and I am very happy with it, plenty snappy and amazing battery life.

 

[geokarbou]

Better title now. Throttling review already done by Notebookcheck for 2017 MacBook m3.

https://www.notebookcheck.net/Apple-MacBook-12-2017-Laptop-Review.230656.0.html

View attachment 708725

Gradually throttles in Cinebench for about 15 minutes, and then crashes through the floor for one run, and then gradually throttles again. However, performance before the crash is on par with the 2016 m5 in your first graph, which is the fastest model in that graph.

I was convinced to get the 2017 but after this review by Notebookcheck, I'm not sure anymore. What I'm saying is that I worry how the Cinebench results translate into real life — under medium workload for an extended period of time. Will it be like the graph? i.e. huge dips in performance with beach balls and then speed back up?

 

[EugW]

I was convinced to get the 2017 but after this review by Notebookcheck, I'm not sure anymore. What I'm saying is that I worry how the Cinebench results translate into real life — under medium workload for an extended period of time. Will it be like the graph? i.e. huge dips in performance with beach balls and then speed back up?

No, not at all. Most usage is idle or low usage with periodic peaks of high usage.

It's not as if you're going to be running Cinebench all day or encoding video all day on a MacBook. Wrong tool for the job.

Yeah, it's pretty amazing that this year's base model is pretty much just as fast at the top model last year (assuming the m5 last year was the top model based on the test above). Got the 2017 m3 and I am very happy with it, plenty snappy and amazing battery life.

I had initially assumed the 2017 was going to get the m3-7Y30 which Turbo boosts to 2.6 GHz, so I was all prepared to buy the i5. But then I found out it was going to use the new m3-7Y32, which is big improvement, and only slightly slower than the i5, so I ordered the m3 instead. It worked out perfectly, since I don't need the 512 GB SSD the i5 comes with. My config is m3 / 256 GB SSD / 16 GB RAM.

Interestingly, last year Apple was shipping out the m5 to reviewers. This year Apple is shipping out the m3 to reviewers.

 

[KPOM]

I was convinced to get the 2017 but after this review by Notebookcheck, I'm not sure anymore. What I'm saying is that I worry how the Cinebench results translate into real life — under medium workload for an extended period of time. Will it be like the graph? i.e. huge dips in performance with beach balls and then speed back up?

According to the Notebookcheck article it takes about 15 minutes of sustained usage to get the throttling. That's a long time.

Remember that Kaby Lake is a refinement of Skylake. GHz for GHz it and Skylake perform the same. Intel was able to increase the clock speed as they have optimized the manufacturing process and improved yields. Because it tops out at higher speeds it may throttle down from them a bit more, but is still likely to perform better than Skylake most of the time. If you already have a Skylake model it is probably not worth the expense to upgrade. However, if you are worried about throttling from everyday tasks I would not be concerned unless you are really taxing the processor (e.g. running Final Cut X for significant editing).

 

[geokarbou]

No, not at all. Most usage is idle or low usage with periodic peaks of high usage.

It's not as if you're going to be running Cinebench all day or encoding video all day on a MacBook. Wrong tool for the job.

According to the Notebookcheck article it takes about 15 minutes of sustained usage to get the throttling. That's a long time.

Remember that Kaby Lake is a refinement of Skylake. GHz for GHz it and Skylake perform the same. Intel was able to increase the clock speed as they have optimized the manufacturing process and improved yields. Because it tops out at higher speeds it may throttle down from them a bit more, but is still likely to perform better than Skylake most of the time. If you already have a Skylake model it is probably not worth the expense to upgrade. However, if you are worried about throttling from everyday tasks I would not be concerned unless you are really taxing the processor (e.g. running Final Cut X for significant editing).

I know, but 15 mins of continuous work is not something rare. I don't do any video or audio encoding which is the most taxing task. My workload consists 90% of browser based apps (on Chrome) which is a RAM hog that's why I'm going to get the 16GB RAM version for sure and most likely the i5 version.

Currently, I have a max specced 13' 2015 MBP (i7, 16GB, 512GB) and the fan starts spinning some times (albeit rarely) because believe it or not, working on Chrome can get demanding CPU-wise!

I do love the weight and portability of the 12' Macbook and weirdly enough, I prefer its keyboard!

I just don't want to pay £1,700 and then see beach balls :/

 

[EugW]

I know, but 15 mins of continuous work is not something rare. I don't do any video or audio encoding which is the most taxing task. My workload consists 90% of browser based apps (on Chrome) which is a RAM hog that's why I'm going to get the 16GB RAM version for sure and most likely the i5 version.

Currently, I have a max specced 13' 2015 MBP (i7, 16GB, 512GB) and the fan starts spinning some times (albeit rarely) because believe it or not, working on Chrome can get demanding CPU-wise!

I do love the weight and portability of the 12' Macbook and weirdly enough, I prefer its keyboard!

I just don't want to pay £1,700 and then see beach balls :/

Chrome can have some moderate usage if you have running advertisements in the browser, but as mentioned, the vast majority of browsing is actually idle time.

The problem in your scenario is that the i7 can run quite hot even with moderate usage.

 

[geokarbou]

Chrome can have some moderate usage if you have running advertisements in the browser, but as mentioned, the vast majority of browsing is actually idle time.

The problem in your scenario is that the i7 can run quite hot even with moderate usage.

I see, thank you.

So, in your experience you don't foresee any issues with my workload?

I really like the 2017 Macbook!