Including Xcode Version in the table may be better ;)

This is more of a discussion question, but do you have any more statistics on how the 16" MacBook Pro, 2019 performed durning the compilation test? For instance, I would like to know how much RAM was used up, and CPU statistics.

Fix Eidolon version

AFAIK, Eidolon is in active development. Specifying only project name you are implicitly saying that you are using master branch, which can be changed.

So, I think it would be better to specify exact commit or tag to ensure tests repeatability.

Simulator

I believe there may be some differences between starting simulator and app on iPad Pro and iPad 2.

Taking into account that the testing project is under active development I'd imagine its complexity is increasing - maybe new dependencies are being added etc., therefore I think it would be good to add date of tests or version of the project under which the performance test had been done.

Alternative consideration: or maybe instead, you could embed buildable version of the project in this repo and use that specific commit for tests so that everyone is testing against same configuration.

Thanks for your work!

Hey Ash,

Is the eidolon repo still the benchmark project for contributors to use?

Just got the 16-inch MacBook Pro and want to be able contribute to this repo by sharing some performance specs across different Xcode versions.

-- Alex

Hey,

Thanks for throwing this together. When you compiled, did you try out the New Build System, and if so did you see any speed performances?

I'm hoping the new build system is better when more cores are available.

Stay awesome!

Geekbench has know relative measurements of different Macs. What could be really cool is to chart these measurements against our measurements to find a correlation between the two. Then we can predict build times for any given Mac based on its Geekbench score.

I'll see if I can get around to this on the weekend.

There appears to be a duplicate entry in the Readme.md for the Mac Mini, Late 2018, 512GB SSD under the X Code 10 table.

It's way too narrow when rendered on the main repo page. We should remove things that are redundant or too verbose:

• "-inch" can be the ".
• We don't need to specify "Intel" anymore 😛
• Retina can be symbolized with the r prefix.

So this:

MacBook (Retina, 12-inch, Early 2015) | 1.1 GHz Intel Core M

Can become:

12" rMacBook Early 2015 | 1.1 GHz Core M

And we can hopefully make things more legible.

Curious to know what configuration would be optimal for Xcode on the new Mac Pro. For example, would the 28-core configuration yield any benefit compared to the 16-core configuration?

I wanted the results sorted by build time so I created a google sheet with sorting. I decided to post it here in case anyone else is also interested. Cheers!

https://docs.google.com/spreadsheets/d/1iYbkv9IIcBMkTxp1vud1Vwjqw_uwMY8kvHIDUTEl9H8/edit?usp=sharing

First thank you for organizing this collective effort.
I want to ask that if we're using the build time with ⌘+B showed by Xcode by sudo defaults write com.apple.dt.Xcode ShowBuildOperationDuration -bool YES ? I think this is a better bench compare to ⌘+R because people might be launching different simulators(for example I am running iPad pro 21.9 3rd Gen). I think different simulators might launch in different timeframe. Also the Xcode build time shows more accuracy, to the 3rd decimals , compare to a stopwatch. For example my build time is 16.641s, 17.478s in 2 tests, compare to when app shows in simulator which is around 27s with stopwatch.
I will submit a pr with the correct number after your response.

hey guys can we add all new mac mini please

... which was introduced by #129.

Edit: I guess the date is also wrong.

APFS can be up to 40% slower than HFS+, so the compile time difference between Sierra and High Sierra should be very significant, it would be nice to know which benchmarks were done on which OS.

I'm really wondering how much faster are the newly refreshed MacBook Pros really are.
Would be nice to have them listed here!

The README should mention which simulator should be run (iPad Air, iPad Pro, iPad Air 2, etc) in case there are differences in launch times among them.

Wanted to post results of my two macs, but could get the test project to build because of this issue.

Is this really the best project to do benchmarking on as it requires so many dependencies which can lead to multiple issues? Also, the eidolon project is constantly updated, which can affect build times making results incomparable.

My suggestion would be to add a test project to this repo which doesn't use any 3rd party dependency, is easy to setup and is never changed. This would probably lead to more contributors.

i guess can run in 35 seconds like the 9900k ？

if i do

rm -rf ~/Library/Developer/Xcode/DerivedData/

it will cost more time.
so I think should cmd+shift+k ,and remove DerivedData will more accurate.