Intel Macs performance vs M1 Macs

[Danny1982]

Initial launches of Rosetta apps takes longer because Rosetta is finishing up the recompiling of the code. Once that has completed, subsequent launches of those apps are much faster. That is something that has been open knowledge since WWDC. There is no "marketing trick", and not "all early video reviews" were conducted improperly, because if they only tracked initial load times instead of normal (post installation and conversion) load times, everything would be skewed for non-native apps. Additionally, that performance penalty only applies to the initial launch of Intel apps, not native apps.

Yes but I think that @RigSatMe talking about Apple native apps. Or not?

 

[RigSatMe]

Initial launches of Rosetta apps takes longer because Rosetta is finishing up the recompiling of the code. Once that has completed, subsequent launches of those apps are much faster. That is something that has been open knowledge since WWDC. There is no "marketing trick", and not "all early video reviews" were conducted improperly, because if they only tracked initial load times instead of normal (post installation and conversion) load times, everything would be skewed for non-native apps. Additionally, that performance penalty only applies to the initial launch of Intel apps, not native apps.

Sorry for misunderstanding, I was referring only to Apple core apps like: Pages, Numbers, Keynote...

 

[RigSatMe]

Yes but I think that @RigSatMe talking about Apple native apps. Or not?

Yes, native apps only!

 

[jeanlain]

Initial app launch speed is mostly limited by SSD speed, which isn't radically better on M1 Macs.

 

[AAPLGeek]

Initial app launch speed is mostly limited by SSD speed, which isn't radically better on M1 Macs.

M1 Air's SSD speeds are twice as fast the 2020 Intel Air.

 

[jeanlain]

M1 Air's SSD speeds are twice as fast the 2020 Intel Air.

At sequential read/write. I haven't seen comparison for random read/write speed.

 

[James_C]

At sequential read/write. I haven't seen comparison for random read/write speed.

You will not see that for SSD's as it makes no difference - that is for conventional HDD where the arm has to track across the disk.

 

[jeanlain]

You will not see that for SSD's as it makes no difference - that is for conventional HDD where the arm has to track across the disk.

?
There are tons of tests for random read/write speed on SSDs, at various block sizes, and speeds can be much lower than for sequential tasks.

 

[James_C]

There are tons of tests for random read/write speed on SSDs, at various block sizes, and speeds can be much lower than for sequential tasks.

Well I learnt something today. You are correct and was not. There is a huge difference between random Read/Write speeds between HDD (around 10ms ) and SSD's (<1ms), which is why I did not think it was relevant to SSD. However once you get past that it is true that different SSD's have difference performance on random read/write.

 

[MBHockey]

Unfortunately, all early video reviews have been not properly carried out in terms of apps opening speed. Did it myself at the Apple store. Apps opening speed is way slower when open apps for the first time. For instance, Numbers bouncing several times before it opens (around 5 times in my experience at MBA M1). Definitely, once the app is cached, it opens right away, but the same happens on Intel. So it’s just marketing trick.

you have no idea what you’re talking about

 

[RigSatMe]

Might be, but as the end user, I care about the fact in the real world practice, where one of the highlighted advantage of M1 on MBA and MBPro is just a gimmick.

 

[AAPLGeek]

At sequential read/write. I haven't seen comparison for random read/write speed.

Here:

It’s true that random read/write speeds affect majority of computing tasks including cold boot and app launching. However, my ancient 2012 MacBook Pro with a crappy SATA SSD boots up in 14 seconds flat and app launches rarely ever take more than a since bounce.

There’s simply no reason for a modern Mac with a crazy fast NVMe drive to be slower than that. Unless it’s due to poorly optimised software which I’m pretty sure is the problem here.

 

[Ethosik]

Hi all.

I read everywhere how the performance of the new M1 vs Intel was amazing (even vs I9 model!!).
So, obviously the support to a new architecture it is not developed in a year, so my question is: and if Apple was working on the optimization of the new architecture already before big sur at the expense of the optimization of the intel chips?

I think that examples like this (founded online): "Developer Paul Hudson shared this example of his M1 MacBook Pro with 16GB RAM rip through an Xcode unzip in just 5 minutes, meanwhile his 8-Core Intel i9 16-inch MacBook Pro with 64GB RAM took over 13 minutes to do the same." can only be explained by the optimization of the features of the new chip.

What do you think about?

Its a problem with Intel. People sometimes say Apple should use AMD instead but AMD came out of nowhere with their amazing processors, long before Apple set this transition in motion at the corporate level. And Apple can make their SoC better optimized than either Intel or AMD by having custom components that do certain tasks that Intel and AMD cannot implement. Like the 16 core Neural Engine for example.