Xcode performance on Macbook 2017 and remote iOS development

[metalsquid]

Hey, is anyone doing iOS development out there on a 2017 rMB 12" and could share your experience? I'm trying to decide between the m3 model and the base 2017 13" MBP nTB to be my sole, daily driver. Either one would be upgraded to 16GB RAM coz I'm already over 2GB swap file with just 8GB.

I've seen on YouTube the new rMB handles 4k and 8k video editing in FCPX with ease but I know that software has been highly optimised. I'm wondering if Apple also did the same voodoo on XC9 and High Sierra that would let it also handle being the daily workhorse for iOS Swift development. Geekbench 4 shows nearly identical scores for the m3-7Y32 and i5-5257U in the 2015 13" MBP. I've used that MBP model as my sole dev machine before and it was fine for my workload. So if I could get identical real world performance in a rMB package that would be fantastic. I'm willing to trade off the pros of the 2017 MBP for more portability, as long as I could be certain that it can handle the workload.

I've thought about a 2012 Mac Mini i7 and 12.5" iPP combo. I do have a fixed IP at home so I could use a Remote Desktop app from the iPP to control XC on the Mini. But if something happens on the Mini end for eg. it needs a reboot, then that undermines the ability to work remotely. I can't just fly back from another country to reboot it. I wish we had Xcode on iOS already...

 

[Greene]

I used to own a 2015 macbook 1.2 (much less power than the current machines), and it handled Xcode fine. Your build times will be a little longer, but not terrible, especially if you're not building all the time.

Honestly though, I would go for the 13", or a 2015 15". Not so much for power reasons, but for screen size. When the 2015 macbook pros and macbooks came out, this guy had a pretty good video on their difference from a developers perspective.

 

[rmm805]

I agree. I had the 2015 MacBook and had no issues with Xcode performance. I don’t think you’re going to do your development on an iPad Pro via Remote Desktop. I think the lag plus the lack of mouse control would be annoying. My iPad Pro 12.9 was fine for Playgrounds but I don’t think I’d want to actually do an iOS app on it (unless you don’t use storyboards).

 

[metalsquid]

I used to own a 2015 macbook 1.2 (much less power than the current machines), and it handled Xcode fine. Your build times will be a little longer, but not terrible, especially if you're not building all the time.

Honestly though, I would go for the 13", or a 2015 15". Not so much for power reasons, but for screen size. When the 2015 macbook pros and macbooks came out, this guy had a pretty good video on their difference from a developers perspective.

Thanks, that was one of the videos I had came across in researching the topic and what got me excited when he mentions that Geekbench multi scores are a good indicator for comparing build times between machines. So I'd expect the MB 2017 to perform the same as the MBP 2015 since their GB scores are pretty much identical. And the MBP 2015 was plenty performant enough for my usage. Definitely a big step up from the 2010 MBA I had been using up until a year ago.

Screen size isn't a big issue for me. I'm used to working on the MBA res and when not mobile I connect to an external display. When I'm traveling between countries or commuting to co-working spaces I really value the smaller and lighter form factor in my backpack. The 15" is definitely out of the question. When I got the 13" MBP the extra weight was noticeable and I didn't like the uniform body thickness. I longed for the old MBA form factor so when I needed to free up some cash I sold it and traded down to a used 2014 MBA. Ironically, after using the MBP for just a few short months I had subconsciously gotten used to the retina display and force tough trackpad, so much so the MBA now feels ancient. Now I'm able to afford to upgrade again, the thought of moving to the MB form factor excites me as it's even thinner, lighter and more modern than the MBA yet will still have enough juice for my needs.

I agree. I had the 2015 MacBook and had no issues with Xcode performance. I don’t think you’re going to do your development on an iPad Pro via Remote Desktop. I think the lag plus the lack of mouse control would be annoying. My iPad Pro 12.9 was fine for Playgrounds but I don’t think I’d want to actually do an iOS app on it (unless you don’t use storyboards).

Yeah, you're probably right. I had the first gen 12.9" and tried to make it my sole device once. Having to reach up and touch the screen all the time got old real fast. I couldn't justify keeping it over the MBA so I sold it. With iOS 11, updated specs and improved apps I was hoping it would be possible at the end of 2017 but looks like it still isn't. The biggest issue is just the inability to use a mouse or trackpad with the iPad. Also, sometimes there are situations where you don't have wifi or 3G service then you're SOL. At least with a MB you can still work offline and build with XC.

Well, looks like I have my answer then, thanks guys!

 

[robvas]

Thanks, that was one of the videos I had came across in researching the topic and what got me excited when he mentions that Geekbench multi scores are a good indicator for comparing build times between machines. So I'd expect the MB 2017 to perform the same as the MBP 2015 since their GB scores are pretty much identical. And the MBP 2015 was plenty performant enough for my usage.

I would not take that as an indicator of compile/build performance.

I have two Macs, a 2016 MacBook 12" and a 2014 MacBook Pro 15", I do part-time development work.

Sure, the 12" MacBook can run all my tools just fine. However, the 15" is easily twice as fast. Probably 4 times faster on things like installing a new version of Xcode or a long build. If I was doing this full-time or could only have one machine, there's no way I would choose the 12".

 

[metalsquid]

I would not take that as an indicator of compile/build performance.

I have two Macs, a 2016 MacBook 12" and a 2014 MacBook Pro 15", I do part-time development work.

Sure, the 12" MacBook can run all my tools just fine. However, the 15" is easily twice as fast. Probably 4 times faster on things like installing a new version of Xcode or a long build. If I was doing this full-time or could only have one machine, there's no way I would choose the 12".

Xcode build time does benefit from the extra cores so naturally a quad-core 15" MBP will run rings around a lowly dual-core rMB. This is also borne out by the GeekBench 4 Multi-Core scores. Assuming you have the base models:

2014 15" MBP - 12572
2016 12" m3 - 5159

However, I was comparing between two dual-core machines, a 13" MBP 2015 and 2017 rMB. Performance should be more equal between these two machines at short interval tasks.

2015 13" MBP - 6721 <- worked on this machine before
2017 12" m3 - 6705
2014 13" Air - 5120 <- my current machine

A clean build on my current machine for a typical project of mine takes about 2m30s. Archive and upload to TestFlight about the same. Incremental builds generally take 30s to < 1min. Quick enough where I hope I won't have to worry about the m3 throttling down too much. Even if it does throttle down to the same performance as I currently get with my current MBA then performance-wise I haven't lost anything from switching but I'd have re-gained a better display and touchpad and gained a much more portable machine. Besides, now is not a good time to be buying a 13" Pro when a refresh with quad-core 8th Gen processors is around the corner.

I'm a full-time developer and the 2014 Air is my one and only machine at the moment. Up until last November I was surviving with a five-year old 2010 Air. I'm lucky enough that I get to work remotely from anywhere so portability for travel is much more important to me than power. More is nice but I can get by with "just enough". I don't need a 15" monster working from a hammock on a tropical beach, or tiny overnight train berths, budget airline seats, etc. iPad Pro would be ideal if I could make it work but sadly I'll still have to rule it out as a sole device for now. Next best option is the rMB. I'll evaluate it for a couple of weeks. If it turns out to be worse than the MBA I'll return it and make do until refreshed 13" MBPs with quad-core are out and either get one of those or a refurb dual-core 13" Pro.

 

[EugW]

I would not take that as an indicator of compile/build performance.

I have two Macs, a 2016 MacBook 12" and a 2014 MacBook Pro 15", I do part-time development work.

Sure, the 12" MacBook can run all my tools just fine. However, the 15" is easily twice as fast. Probably 4 times faster on things like installing a new version of Xcode or a long build. If I was doing this full-time or could only have one machine, there's no way I would choose the 12".

I'm not a developer but...

Besides the fact that as mentioned the MBP15" is quad-core, it should also be noted that the 2017 12" has a significantly faster SSD than the 2016 12". I think the more realistic comparison is the 2017 12" vs an older dual-core 13" MacBook Pro or MacBook Air. Yes, the quad 15" will be way faster than a 2017 12", but it will be way faster than dual-core MacBook Pros and MacBook Airs too.

 

[robvas]

My old 2013 13" MacBook Pro was also much faster than the 12" at things like running an Xcode update, running NetBeans, compiling a big project... Ran them side by side quite a bit just to test. I got rid of the 13" because I got the 12", I still had the 15" if I wanted something fast.

Running an OS X update seriously can take 20-25 minutes longer on the little guy. I think a big part of it is the Pro has fans that will kick on when you are doing something that runs longer than 20 seconds or so, where the 12" will slow down under load.

 

[EugW]

My old 2013 13" MacBook Pro was also much faster than the 12" at things like running an Xcode update, running NetBeans, compiling a big project... Ran them side by side quite a bit just to test. I got rid of the 13" because I got the 12", I still had the 15" if I wanted something fast.

Running an OS X update seriously can take 20-25 minutes longer on the little guy. I think a big part of it is the Pro has fans that will kick on when you are doing something that runs longer than 20 seconds or so, where the 12" will slow down under load.

But as mentioned, the 2017 12" is way faster than the 2016 12". In fact, my 2017 m3 is faster than all three models of the 2016 as tested by Notebookcheck, and the 2017 i5 and i7 are even faster.

Plus the 2016 i7 in their test had issues with throttling, but it might be not be a consistent problem and just a problem with individual i7 machines.

Which 2016 do you have?

FWIW, my 2017 m3 loses about 4-7% CPU speed at near 100% CPU load after about 25 mins, so the speed loss is not huge with sustained usage. The graph below is from the listed thread, where some of us forum members benched our machines.

https://forums.macrumors.com/thread...uns-of-cinebench.2073415/page-2#post-25271927

What you can see here is that an 2017 MacBook i7 runs close to the speed of a 2017 i5 MacBook Air, although both are significantly slower than the 2017 13" i5 MacBook Pro.

https://www.notebookcheck.net/Apple-MacBook-Air-13-2017-Laptop-1-8-GHz-Review.230010.0.html

 

[metalsquid]

My old 2013 13" MacBook Pro was also much faster than the 12" at things like running an Xcode update, running NetBeans, compiling a big project... Ran them side by side quite a bit just to test. I got rid of the 13" because I got the 12", I still had the 15" if I wanted something fast.

Running an OS X update seriously can take 20-25 minutes longer on the little guy. I think a big part of it is the Pro has fans that will kick on when you are doing something that runs longer than 20 seconds or so, where the 12" will slow down under load.

Aye, that's true. Updates take a while and something taking that long could throttle the m3 and make it take even longer. Could always leave updates to run just before bed time though. Hop on the iPad and read some comics while waiting.

 

[bill-p]

If you don't have to run the Simulator, it works decent enough.

If you do have to use the Simulator, then abandon the idea and get a 15" Pro.