MacBook (m3, m5, m7) Performance Compared to Recent Airs and Pros (Chart)

[Leo90]

You're probably better off with a 2016 m3 then. The SSD/RAM/Battery life improvements outweigh the processor. The new m3 is said to match the old m5, so you're not far off of the old m7 anyway.

But at the price of a new m3 I could get a refurbished m7.

 

[justinl01]

But at the price of a new m3 I could get a refurbished m7.

You could but then you would miss out on the far faster SSD I/O access which is significant and probably more noticeable in every day use than the M7 processor would be.

 

[keviig]

You could but then you would miss out on the far faster SSD I/O access which is significant and probably more noticeable in every day use than the M7 processor would be.

My point exactly. Under throttled load the m3 is likely as good, or better than the old m7 as well.

 

[deeddawg]

you would miss out on the far faster SSD I/O access which is significant and probably more noticeable in every day use than the M7 processor would be.

To quantify this, we're talking read speeds of around 950 vs 850 MB/s and write speeds of around 740 vs 450 MB/s in comparing the 2016 to 2015. So a modest gain in read speeds and a pretty good bump in write speeds. https://www.macrumors.com/2016/04/25/2016-macbook-review-roundup/

Which scenarios are you thinking of where this makes enough difference to note?

I'm not coming up with everyday scenarious where this would noticeably impact the user. This is why I'm asking.

It seems you have to be reading/writing multi-gigabyte files to even notice a difference. Further, it seems to me that in most such scenarios you're applying some sort of processing to the files and I'd think other factors (CPU/memory/network) would be the limiting factor before the SSD.

Your thoughts?

 

[izzyfanto]

You're probably better off with a 2016 m3 then. The SSD/RAM/Battery life improvements outweigh the processor. The new m3 is said to match the old m5, so you're not far off of the old m7 anyway.

That's why I bought my m3
[doublepost=1462483042][/doublepost]

To quantify this, we're talking read speeds of around 950 vs 850 MB/s and write speeds of around 740 vs 450 MB/s in comparing the 2016 to 2015. So a modest gain in read speeds and a pretty good bump in write speeds. https://www.macrumors.com/2016/04/25/2016-macbook-review-roundup/

Which scenarios are you thinking of where this makes enough difference to note?

I'm not coming up with everyday scenarious where this would noticeably impact the user. This is why I'm asking.

It seems you have to be reading/writing multi-gigabyte files to even notice a difference. Further, it seems to me that in most such scenarios you're applying some sort of processing to the files and I'd think other factors (CPU/memory/network) would be the limiting factor before the SSD.

Your thoughts?

The SSD speeds weren't that important to me. I got it for the more optimized 6th gen CPU, higher clocked RAM, and better GPU and better battery life.

 

[SchodMC]

To quantify this, we're talking read speeds of around 950 vs 850 MB/s and write speeds of around 740 vs 450 MB/s in comparing the 2016 to 2015. So a modest gain in read speeds and a pretty good bump in write speeds. https://www.macrumors.com/2016/04/25/2016-macbook-review-roundup/

Which scenarios are you thinking of where this makes enough difference to note?

I'm not coming up with everyday scenarious where this would noticeably impact the user. This is why I'm asking.

It seems you have to be reading/writing multi-gigabyte files to even notice a difference. Further, it seems to me that in most such scenarios you're applying some sort of processing to the files and I'd think other factors (CPU/memory/network) would be the limiting factor before the SSD.

Your thoughts?

About the write-times: I think that only will be noticeable when writing a lot of data, like transfering your iTunes library from an other mac, installing an OS update, etc. Even if the there will be written a lot to the SSD while running your system, I think that there will be no significant slowdown in every day usage. A lot of little files will be written fast enough to be fine for daly usage. IMHO.

Read-times: That comes in handy when you boot up your system or start an app for the first time. Starting an app for a second time - without rebooting the system - would be fast because of the memory management of OS X. If you have a look at the activity monitor, memory, you will notice a large block called "inactive" (I think meanwhile they renamed it to "files in cache"). When you close your app, the executable part and some kind of internal data (like icons, windows of the app, etc) will stay in memory as long as that memory won't be used by an other app. When you start the app again, that part cached in memory will be used. Because of this starting an app a second time is much faster. (If they changed that mechanism in one of the last OS updates, please tell me).

Well, will you notice a difference in your daily needs? I don't think so. Maybe you will notice a difference if you directly compare the two models or if you are alrady comfortable with something faster than the 2015 models read / write speed. If you are comming from a normal HDD or a slower SSD, you will be blown away from the SSD speed, anyway (the first months with my 2013 MBA that had an PCI Express SSD for the first time, I was so exited how fast that system boots up and loads apps compared to my 2011 MBP with an upgraded SATA SSD!).

While using your system, you will start an app and save your files. And I have the strong feeling, that you won't sit with a stopwatch in front of your computer to stop the time it needs for that tasks.

So if the overall performance and the battery life of the 2015 Modell is fine for you, you can save money and use it for other usefull stuff like adapters, iTunes content, pay for a candle light dinner with someone special - which could have more long-time value than buying a Macbook, at least I hope that for you. And btw, let your Macbook at home for that time...

 

[deeddawg]

Well, will you notice a difference in your daily needs? I don't think so.

Good post. You echo my thoughts on the matter. Yes, there's a substantial increase in SSD write speeds and a moderate increase in read speeds, but the 2015 starting point is already high enough that the differences in real world usage (for most people) would be minor at best. Still, I'm open minded and curious as to any real scenarios where there'd be a substantial difference.

Lest anyone misunderstand, I do think the 2016 is a decent overall incremental upgrade from the 2015. Enough to give people something to think about in terms of choosing a cheaper refurb or leftover 2015 vs. a full price 2016. Probably not enough for most with a 2015 to upgrade though, but individual situations vary.

 

[zhenya]

The improvement in random read/write speeds will be much more noticeable than the sequential speeds, especially if you are encrypting your drive.

 

[deeddawg]

Interesting that the numbers are so different from the QuickBench small random read and write scores reported for the 2015 rMB over at BF: http://barefeats.com/rmb12in.html

I'd give QB a shot on my 2015 but looks like I'd have to buy it? Not that interested in the matter.

 

[zhenya]

Interesting that the numbers are so different from the QuickBench small random read and write scores reported for the 2015 rMB over at BF: http://barefeats.com/rmb12in.html

I'd give QB a shot on my 2015 but looks like I'd have to buy it? Not that interested in the matter.

I'm not intimately familiar with the details of testing SSD's, but a first-order look shows that Ars' test was done with a fixed 8k block size, while the barefeats review is a random test with block sizes ranging from 4k to 1024k. Given that Ars fixed theirs at nearly the smallest block size, I'd expect their numbers to be significantly lower.

 

[PatriotInvasion]

Shockingly? No. It's slower. For email, web surfing, drafting documents, etc. it's totally fine. For Photos for Mac edits and organization, it's good but I imagine the m5 or m7 would be a bit better.

For travel, it's better for my use than either my former 13" rMBP or trying to pretend my iPad Air is a netbook.

So I couldn't resist the $999 sale price on Amazon for the Early 2015 base MacBook. Received mine today and couldn't be happier with this machine thus far as a compliment to my iMac. It was hard for me to stomach a secondary Mac at the $1,299 price point, but at $999 it was a no brainer for me. My use case when away from my desk makes the MacBook more than sufficient since any video editing or extensive photo editing I'd prefer to do on my 27-inch iMac anyway.

I think anyone else in my shoes should really consider hopping on Amazon and taking advantage of the $300 discount on Early 2015 models asap while they last.

 

[c0ppo]

Don't really play too much with Firefox. But I doubt the Google Chrome store extensions would be compatible.

Then install Vivaldi at least. Sure, memory + cpu hog as is Chrome, but way more customizable, has tab stacking, sessions management and a million of other things chrome can only dream about. And can use chrome extensions, and even opera extensions

 

[Leon88]

Quick update. So I was processing about 30 odd camera raw files (Leica M9) with the m7 MB yesterday. Whilst I didn't notice any difference while editing my photos (slider adjustments/spot editing/filters) when compared against my Late 2013 13" MBP, I did notice that batch exporting all 30 images took much longer than on the MBP. My MB started getting pretty warm too- I think throttling was starting to kick in

 

[Trey M]

IMO the the real thing you're getting with a 2016 model is a better optimized processor and an extra hour of battery life. To me, it's as simple as that. I can't imagine buying a 2016 model over a 2015 for a slightly faster SSD and very very slightly improved RAM speeds (like 266 MHz). If a better optimized processor (modest clock upgrade, smarter throttling) and extra battery life are valuable to you, opt for 2016. If not, find a great deal on the internet for a 2015 model.
[doublepost=1462930354][/doublepost]

Quick update. So I was processing about 30 odd camera raw files (Leica M9) with the m7 MB yesterday. Whilst I didn't notice any difference while editing my photos (slider adjustments/spot editing/filters) when compared against my Late 2013 13" MBP, I did notice that batch exporting all 30 images took much longer than on the MBP. My MB started getting pretty warm too- I think throttling was starting to kick in

In the short time I've spent with my new rMB I've noticed it getting fairly hot as well. Hard to necessarily tell if it was throttling or just shedding heat. I noticed the first day and a half the thing was really, really sluggish, but from past experience I think that's just the system and Spotlight indexing the SSD.

 

[TigerMSTR]

I wish Apple would have Spotlight index when the Mac is sleeping and plugged into charge, rather than when you're using it. Every time there's a new version of OS X the Mac becomes unbearably slow as it indexes, even on my MacBook Pro.

 

[NavySEAL6]

I wish Apple would have Spotlight index when the Mac is sleeping and plugged into charge, rather than when you're using it. Every time there's a new version of OS X the Mac becomes unbearably slow as it indexes, even on my MacBook Pro.

Newer Macs with Powernap can do this

 

[PatriotInvasion]

No, not painful at all. Sorry if it sounded that way.

You only need to mess with symlinks if you want to automatically handle standard default folders that apps may just default to such as ~/Documents as one example. Otherwise just put your GoogleDrive or CloudStation folder in the Favorites list and you're done.

The syncing stuff is going to apply to most anything you're trying to span across multiple computers. If your rMB is shut down and you've made a bunch of files changes on your iMac, then head out for your flight and start to edit the stuff on your rMB without syncing you're risking a mess. Especially with something complex like a Lightroom photo library folder structure.

I think I have a much easier solution...Back to My Mac and iCloud Drive. Basically, I just keep every regular document in iCloud Drive and that syncs automatically across Macs. For stuff I have on my iMac and not in iCloud Drive, using Back to My Mac works perfectly on my 12-inch MacBook. I just access the file that way as needed and save.

 

[jamrolu]

I've had an m3 and m7 side-by-side all this week trying to decide which I want. I've come to a few conclusions from using the Intel Power Gadget.

• The m7 runs hotter, the temperature rockets when boosted but takes longer to fall back, and never reaches the low temperature of the m3.
• No noticeable difference in UI responsiveness.
• Peak power usage on the m3 is around 7W, and 14W for the m7.
• The battery life seems marginally better on the m3 (not scientifically quantified!).

I put down a few more observations and screenshots in this thread https://forums.macrumors.com/threads/macbook-1-1ghz-m3-vs-1-3ghz-m7.1971827/

 

[zhenya]

I've had an m3 and m7 side-by-side all this week trying to decide which I want. I've come to a few conclusions from using the Intel Power Gadget.

• The m7 runs hotter, the temperature rockets when boosted but takes longer to fall back, and never reaches the low temperature of the m3.
• No noticeable difference in UI responsiveness.
• Peak power usage on the m3 is around 7W, and 14W for the m7.
• The battery life seems marginally better on the m3 (not scientifically quantified!).

I put down a few more observations and screenshots in this thread https://forums.macrumors.com/threads/macbook-1-1ghz-m3-vs-1-3ghz-m7.1971827/

That doesn't make sense. Both should be limited to the same power envelope, and much closer to 7W than 14.

 

[jamrolu]

That doesn't make sense. Both should be limited to the same power envelope, and much closer to 7W than 14.

Then there must be a bug in the Intel Power Gadget. Here are the screenshots from the other thread, which was while opening Lightroom. m3 on the left, m7 on the right:

 

[zhenya]

Then there must be a bug in the Intel Power Gadget. Here are the screenshots from the other thread, which was while opening Lightroom. m3 on the left, m7 on the right:

View attachment 631138 View attachment 631140

That has to be a bug. The u-series processors with a fan start at 15w. The various iterations of Core-M are all the same silicon, just sorted after production into different batches. Interesting findings nonetheless! It's great to get some first-hand knowledge that the fastest model doesn't make a lot of practical difference. Interesting that it runs hotter though - I would have expected the opposite.

 

[jamrolu]

That has to be a bug. The u-series processors with a fan start at 15w. The various iterations of Core-M are all the same silicon, just sorted after production into different batches. Interesting findings nonetheless! It's great to get some first-hand knowledge that the fastest model doesn't make a lot of practical difference. Interesting that it runs hotter though - I would have expected the opposite.

You can see the bottom graph is the CPU temperature and it rockets up in sync with the spike in power consumption. The 14W moments are brief but definitely happened whenever opening a 'heavy' app.

 

[deeddawg]

I think I have a much easier solution...Back to My Mac and iCloud Drive.

Except that doesn't solves your desire for a local-network based sync. All your iCloud syncs are round-tripping across the Internet even when all devices are on the same LAN.

Just wish there was a 1st party, local-network based sync system that would merge the file system of two Macs into one despite being on separate machines.

I've achieved that in practice and described how I solved what you said you wanted.

If you don't really want to do what you said you wanted to do, that's fine. I was trying to be helpful, not to convince you to do anything. Just don't mislead others by ignorantly dismissing as difficult something that isn't.

 

[keviig]

That has to be a bug. The u-series processors with a fan start at 15w. The various iterations of Core-M are all the same silicon, just sorted after production into different batches. Interesting findings nonetheless! It's great to get some first-hand knowledge that the fastest model doesn't make a lot of practical difference. Interesting that it runs hotter though - I would have expected the opposite.

That's not a bug. A rMBP needs roughly 20w to hit 3.1ghz turbo speed (as tested with my rMBP 5257u). Roughly 14w at 2.8ghz according to IPG is to be expected. Just because it's a core m doesn't mean it can magically turbo up to high speeds without using more power. Look at how the temperature spikes when comparing the m3 and m7. The m7 goes from right under 40c to 80c in the same (or less) amount of time the m3 goes from just under 40c to 60c, that's double the temperature delta from idle! The m7 is clearly using a whole lot more power than the m3.

 

[jamrolu]

That's not a bug. A rMBP needs roughly 20w to hit 3.1ghz turbo speed (as tested with my rMBP 5257u). Roughly 14w at 2.8ghz according to IPG is to be expected. Just because it's a core m doesn't mean it can magically turbo up to high speeds without using more power. Look at how the temperature spikes when comparing the m3 and m7. The m7 goes from right under 40c to 80c in the same (or less) amount of time the m3 goes from just under 40c to 60c, that's double the temperature delta from idle! The m7 is clearly using a whole lot more power than the m3.

And while it sucked up twice the power it didn't get back to idle in half the time...