1.1 VS 1.2 real tests

[tutux]

Hi guys,

I'll need help to understand how a 1.1 rMB can be faster than a 1.2 (especially because I canceled my 1.1 to get the 1.2 and now seeing that kills me )


Source (in french, sorry): http://www.macg.co/tests/2015/05/test-du-macbook-12-debut-2015-core-m-12-ghz-88989

and the tests results:

Imovie and Garageband are 10% slower on 1.2 (and this is a real usage not a benchmark).

Can this be related to the fact that 1.1 is a 2 weeks old machine (with all setup/indexed etc) and the 1.2 a more fresh machine??

 

[leman]

I think what it shows is that the CPU speed doesn't really matter. And we have no ida about their testing methodology. One should do the test several times and compare spread, not just the average.

 

[tutux]

I think what it shows is that the CPU speed doesn't really matter. And we have no ida about their testing methodology. One should do the test several times and compare spread, not just the average.

I can understand similar tests, but worst result with the fastest... it doesn't compute in my brain


Lets assume they performed the test multiple times to check how reliable it is

 

[Kiwi 99]

Hi guys,

Can this be related to the fact that 1.1 is a 2 weeks old machine (with all setup/indexed etc) and the 1.2 a more fresh machine??

Absolutely it can be. That could easily be the reason. Let the machine settle in for a week and then try again.

(And keep an eye on Activity Monitor while you are testing, to make sure there's no other programs deciding to wake up and do some tidying up of their own)

 

[Narcaz]

I don't think this is surprising. They applied sustained heavy load over 2 minutes and than the Core-M processor starts to throttle. It doesn't sound logical at first, but the faster model probably needs to throttle harder, because it reached the thermal limits first. If the goal is to run 10 km, it started like Usain Bolt, slowed down and needs to recover for a while. Some people around here are to focused on geekbench results and numbers. They think the 1.2 ghz model must always be better than the 1.1 ghz model. They have different turbo states, so the 1.2 looks good when you need a short performance boost (such as Geekbench).

But all Core M processors have the same thermal limits. So after a while performance will degrade and could even fall behind the 1.1 ghz model for a particular task. Even Anandtech noticed this behavior. The fanless Asus Zenbook UX 305 with 0.8ghz Core-M performs better in e.g. DOTA than the 1.1ghz MB, because it doesn't need to throttle so hard:

 

[ZipZap]

I don't think this is surprising. They applied sustained heavy load over 2 minutes and than the Core-M processor starts to throttle. It doesn't sound logical at first, but the faster model probably needs to throttle harder, because it reached the thermal limits first. If the goal is to run 10 km, it started like Usain Bolt, slowed down and needs to recover for a while. Some people around here are to focused on geekbench results and numbers. They think the 1.2 ghz model must always be better than the 1.1 ghz model. They have different turbo states, so the 1.2 looks good when you need a short performance boost (such as Geekbench).

But all Core M processors have the same thermal limits. So after a while performance will degrade and could even fall behind the 1.1 ghz model for a particular task. Even Anandtech noticed this behavior. The fanless Asus Zenbook UX 305 with 0.8ghz Core-M performs better in e.g. DOTA than the 1.1ghz MB, because it doesn't need to throttle so hard:

Image

It is my understanding that the faster cpu is supposed to be more thermally efficient. I would expect to see the faster chip reach the throttling level the slowest.

 

[tutux]

It is my understanding that the faster cpu is supposed to be more thermally efficient. I would expect to see the faster chip reach the throttling level the slowest.

or at least be the same if the constraint are the same

 

[Narcaz]

It is my understanding that the faster cpu is supposed to be more thermally efficient. I would expect to see the faster chip reach the throttling level the slowest.

How? All Core-M are made with same production process, have the same processor architecture and the same TDP (up to 6Watt). They only have different clocks speeds and turbo states. Apple confirmed to a german techblog, that their Core-M processors are even limited to a custom TDP of 5 Watt.(http://www.golem.de/news/macbook-12-teurer-prozessor-kaum-schneller-1504-113366.html).

or at least be the same if the constraint are the same

I think that Speedstepping or switching turbo states is a rather complex task. The higher clocked model is faster for short periods of loads, because it can take advantage of thermal intertia (i hope it's the right term, i am not a physicist nor a native english speaker). When the processor uses its turbo mode, the temperature rises slowly. So it takes a few seconds until it has to reduce its frequency and voltage. In the meantime a task like sharpening a picture is probably done.

If you apply sustained loads there is a downside. The processor will heat up other components like the mainboard, chassis etc. more quickly. This effect influences the processors, so it can't make uses of it's turbo clocks anymore and will reduce its speed. And it probably takes a little bit longer to cool down. The lower clocked model reaches this state a bit later than the higher clocked model, because its turbo is slower. And therefore it can keep up its speed a bit longer and produces the same or slightly better results sometimes. Look at the temperatures and average speed from the french test (started on both machines at the same time during iMovie Export):

I am not saying that Core-M is a crappy processor or underpowered. It's actually very advanced engineering and probably sufficient for most users. But in my opinion old thumb rules (higher clocks = always faster) or geekbench results only tell one side of the story. There is a certain variety in real life tests due to heat sensitivity of the Core M. I couldn't find equivalent in depth english article, where the Core M is tested in the same way. So my posts is based on this article from the german techblog golem: http://www.golem.de/news/core-m-5y70-im-test-vom-turbo-zur-vollbremsung-1411-110529.html

 

[ZipZap]

How? All Core-M are made with same production process, have the same processor architecture and the same TDP (up to 6Watt). They only have different clocks speeds and turbo states. Apple confirmed to a german techblog, that their Core-M processors are even limited to a custom TDP of 5 Watt.(http://www.golem.de/news/macbook-12-teurer-prozessor-kaum-schneller-1504-113366.html).



I think that Speedstepping or switching turbo states is a rather complex task. The higher clocked model is faster for short periods of loads, because it can take advantage of thermal intertia (i hope it's the right term, i am not a physicist nor a native english speaker). When the processor uses its turbo mode, the temperature rises slowly. So it takes a few seconds until it has to reduce its frequency and voltage. In the meantime a task like sharpening a picture is probably done.

If you apply sustained loads there is a downside. The processor will heat up other components like the mainboard, chassis etc. more quickly. This effect influences the processors, so it can't make uses of it's turbo clocks anymore and will reduce its speed. And it probably takes a little bit longer to cool down. The lower clocked model reaches this state a bit later than the higher clocked model, because its turbo is slower. And therefore it can keep up its speed a bit longer and produces the same or slightly better results sometimes. Look at the temperatures and average speed from the french test (started on both machines at the same time during iMovie Export):

Image


I am not saying that Core-M is a crappy processor or underpowered. It's actually very advanced engineering and probably sufficient for most users. But in my opinion old thumb rules (higher clocks = always faster) or geekbench results only tell one side of the story. There is a certain variety in real life tests due to heat sensitivity of the Core M. I couldn't find equivalent in depth english article, where the Core M is tested in the same way. So my posts is based on this article from the german techblog golem: http://www.golem.de/news/core-m-5y70-im-test-vom-turbo-zur-vollbremsung-1411-110529.html

I'll let someone more knowledgable respond. I know I have seen the heat information posted on a number of threads over the weeks, indicating 1.3 was more heat efficient than 1.1 or 1.2

 

[Jetcat3]

How? All Core-M are made with same production process, have the same processor architecture and the same TDP (up to 6Watt). They only have different clocks speeds and turbo states. Apple confirmed to a german techblog, that their Core-M processors are even limited to a custom TDP of 5 Watt.(http://www.golem.de/news/macbook-12-teurer-prozessor-kaum-schneller-1504-113366.html).

Wait, that article states that Apple has limited all three of these Intel Core M's to 5 Watts? Isn't impossible to bump the clock speeds from the base model 5Y31 from .9 to 1.1 without increasing the TDP to 6 watts per Intel's spec sheet? That's impressive if they've kept the base model 5Y31 to 5 watts with the bump in processor speed.

 

[Narcaz]

I'll let someone more knowledgable respond. I know I have seen the heat information posted on a number of threads over the weeks, indicating 1.3 was more heat efficient than 1.1 or 1.2

Like this? https://forums.macrumors.com/threads/1875399/

This side by side real life tests seem to contradict the assumption that the 1.2 will perform better in every usage case. There isn't an explanation for the results from the french blog, anandtech or the german site. It is more than unlikely all them messed up their testing methodology (especially golem has a good reputation). And to highlight it again: the 1.2/1.3 ghz is still faster for short periods of time, which is beneficial for most usage cases of the Macbook. I don't think anyone plans to use this machine 24/7 for gaming, rendering or exporting movies. So the sustained load performance isn't really a big issue.

Wait, that article states that Apple has limited all three of these Intel Core M's to 5 Watts? Isn't impossible to bump the clock speeds from the base model 5Y31 from .9 to 1.1 without increasing the TDP to 6 watts per Intel's spec sheet? That's impressive if they've kept the base model 5Y31 to 5 watts with the bump in processor speed.

The statement doesn't leaves much room for interpretation. Golem discussed in the article the possibiblity of using 4,5 Watt TDP for the 1,1 and 1,2 and 6 Watt for the 1.3. Apple's answer was their Core-M custom TDP is 5 Watt. It isn't very specific, but this seems to apply for all processors of the Macbook.

The public spec sheets of Intel could suggest the use of 6 Watt, but these sheets are for standard models. Apple is using non standard base clocks on these processors. It wouldn't be first time Apple got some cherrypicked special version (probably for cooling reasons). But one knows for sure. It's just speculation.

 

[tutux]

Like this? https://forums.macrumors.com/threads/1875399/

This side by side real life tests seem to contradict the assumption that the 1.2 will perform better in every usage case. There isn't an explanation for the results from the french blog, anandtech or the german site. It is more than unlikely all them messed up their testing methodology (especially golem has a good reputation). And to highlight it again: the 1.2/1.3 ghz is still faster for short periods of time, which is beneficial for most usage cases of the Macbook. I don't think anyone plans to use this machine 24/7 for gaming, rendering or exporting movies. So the sustained load performance isn't really a big issue.



The statement doesn't leaves much room for interpretation. Golem discussed in the article the possibiblity of using 4,5 Watt TDP for the 1,1 and 1,2 and 6 Watt for the 1.3. Apple's answer was their Core-M custom TDP is 5 Watt. It isn't very specific, but this seems to apply for all processors of the Macbook.

The public spec sheets of Intel could suggest the use of 6 Watt, but these sheets are for standard models. Apple is using non standard base clocks on these processors. It wouldn't be first time Apple got some cherrypicked special version (probably for cooling reasons). But one knows for sure. It's just speculation.

Core M processors are customizable, you choose the power, and intel tells you the speed you can get.

4.5W base is 900 / 1100 / 1200 for the 3 models
6W (max) is 1100/ 1300 / 1400 for the 3 models

So the 1.1 is maxed out, and not the 2 others.

 

[pookitoo]

So, what is the conclusion ? 1,1 have the same performance than the 1,2 ? And the 1.3 option is not a good choice ? And benchmark say nothing accurate ? ( 1.1 is similar in benchmark than the MacBook Air 2011, and the 1.3 is near the 2015 MacBook Air , so I was thinking that the upgrade worst it ? I'm confuse )

I' ve one week to send back the 1.3 version and take the 1.1 version... What to do ?

 

[bcaslis]

Keep the 1.3GHz. There was an earlier thread by someone with side to side testing of the 1.2 and 1.3.

 

[pookitoo]

Yes, and i've read it, but the thread talk about the battery life, and don't talk about performance

 

[bcaslis]

No it does talk about performance. Basically real world you aren't going to see a difference but the 1.3 can burst a task faster so it should finish faster and therefore use less battery.

Look at the chart at the top of this thread. The worst case was the 1.1 was faster by 5 seconds for a three minute task, yet in other tasks it was slower. Are the tasks more CPU bound or disk bound. If disk bound than a slightly faster SSD will give a better result.

I got a 1.3 yesterday versus using a 1.2 the past three weeks. According to the Black Magic disk speed test and Geekbench 3 it is slightly faster. Slightly. Do I feel it. Maybe a hair faster. A hair. It's not a big deal but I see no negative side to it. If you can afford the faster processor and are willing to wait to get it, you should. To me the 1.1 is non-starter because the disk is too small.