Docker on M1 Pro and M1 Max | against M1 and Intel

Testing Docker on M1 Max, M1 Pro, Intel Core i9 MacBook Pros and the good old M1 MacBook Air.

CPU temperatures and fan speed readings via TG Pro: (affiliate link)

Related Docker videos:

▶️ Docker on M1 MacBook vs Ryzen 7 Lenovo –
▶️ M1 MacBook SSD and Docker –
▶️ Apple M1 vs Intel Core i9 and Docker Preview – First Look –
▶️ M1 With Docker Live Stream –
▶️ Apple M1 and VSCode Performance –

💻NativeScript training courses —
(Take 15% off any premium NativeScript course by using the coupon code YT2020)
👕👚iScriptNative Gear –

— — — — — — — — —


Click here to subscribe:

— — — — — — — — —


NativeScript Core Getting Started Guide (Free Course) –

NativeScript with Angular Getting Started Guide (Free Course) –

Upgrading Cordova Applications to NativeScript (Free Course) –

— — — — — — — — —



This post was created with our nice and easy submission form. Create your post!

Share this:

What do you think?

72 points
Upvote Downvote


  1. (Docker employee here)
    Bit surprising results on the build side. I looked into it, and the issue seems to be that when you run "yarn install" it installs sqlite3 that has C dependencies. For this package in x86_64 yarn will download the prebuilt sqlite3 library directly, but for arm64 no such prebuilt version exists so it needs to compile it from source code on the machine instead. Hopefully, as these machines become more popular, package authors will start uploading the prebuilt version for arm as well.

    So users might definitely hit this issue as even sqlite3 itself is a very common package but don't take it as a reflection of raw container build performance on M1.

  2. How does docker affect battery life on M1 max Vs M1 Pro Vs M1? Is there a real benefit to having the 14 over the 16 with the M1 Pro or max? These are the things I want to know. The 14 is so much lighter and portable in comparison to the 16. Great to see it's so fast on power.

  3. Your Vid shows to be really in front off the M1 chip , the pro and the max would need 12 or even 16 cpu cores.

  4. It shouldn't be lost on anyone that Apple still hasn't implemented mutithreading (hyperthreading in Intel parlance) in the M1 scalar cores. That's not an oversight or dumb mistake. Hyperthreading is a dumb mistake on Intel's part. I have many Intel microarchitects as friends, and they all say they tried hard to get Intel to drop hyperthreading to no avail (marketing has ruled at Intel far, far too long). It actually hurts overall performance and increases power dissipation. It's amazing that AMD has been able to do a good job with their implementation, but take note of the fact that Intel efficiency cores in their latest processors do not implement hyperthreading. It's because it wouldn't help and would only increase power dissipation. They're learning….slowly.

  5. I do like all of these tests, but to be honest Apple has been known to poorly cool it's intel macs for years and on top of that they often selected last Gen CPU's to put into the machines. So what you are probably getting here is thermall throttling on the intel based macs which will be hurting your performance here.

    Now I am not saying that M1's aren't good (anyone that has used a RISK machine from the 90's or a Raspberry Pi knows that ARM CPU's are good) but I what I am saying is that this is not going to be a fair comparison for out right speed. However, for the average consumer this should testing will be great!

  6. Any reason why you're limiting the memory and CPU cores of the M1 Pro and M1 Max? I mean we know the CPU cores are the same, so the whole point is to get additional cores and additional RAM over the 8 cores and 16 GB of RAM you get with the M1 Macbook Pro to speed things up => a lot of people in the comments that are getting confused by your testing methodology. Artificially limiting the performance of the M1 Pro and M1 Max isn't doing anyone any favors here.

  7. Thanks for the effort.

    This seems to demonstrate how the 14"/16" enclosures with fans would help avoid thermal throttling – even when they are constrained to match the M1 Air (cores, RAM).
    Wonder how those beasts perform when completely unleashed – 100% of CPU+RAM as configured in M1 Air.

  8. I love the idea of this test, but I wish there were more knowledge about docker. Half of the tests docker ended up failing and bailing early. The M1 Air wasn’t taking 50+ seconds, docker itself was getting (probably) OOM killed or watchdog killed and the timer was just waiting for the timeout

  9. Why in the world you are limiting the docker configs??? What people are interested is to see how much more powerfull those 10 core cpus are and how much they affect the day to day life of developers. I am not interested in fair fight, I am interested in how much more those 2 extra cores are worth to me.

  10. So when using yarn, some packages can be pre-compiled for x86, while they are rarely pre-compiled for ARM. So when x86 downloads packages, ARM might take some time for compilation under the hood.

  11. And now do the same thing with a Thinkpad running Linux or even Windows. My Desktop i7 10700 did not break a sweat, did not even use 4 GB of RAM (25 per container). Docker is still starting up a virtual machine on macOS and this so resource hungry.

  12. Hello sir, I think you should try these test without screen recording on M1, M1 Pro and M1 Max.

  13. all servers are still x86_64, so everyone how work with docker need to work with x86_64 images not arm64.

  14. I am new to mac world here. What app is used to show the temp, fan and cpu for mac? I see a lot of folks have it but I don't know what the app is.

  15. I'm getting faster speeds using the new option in silicon Docker Deskop to use Big Sur virtualization (instead of hypervisor), in Experimental Features settings. Also using option for gRPC FUSE and seeing faster speeds from host to container file access. Drupal runs fast even with xdebug on.

  16. I just want to say thank you for doing these developer relevant tests. I was tired of geekbench scores and video encoding tests

  17. I am looking to see how many VMs can we run on M1 Max 64GB). My application is to create a portable lab on M1 Max. Usually run Linux servers (Web, NFS, SFTP, SMB, Open Virtual Switch, syslog), windows servers (AD, MSDB, EPO server), windows workstations, load balances (F5 VM) and BSD firewalls or palo alto firewalls.

  18. I really do like the comparisons but please take the time to be a bit less sloppy (like just ignoring the build failure on the intel, arbitrarily deciding that 16 cores with HT will be identical to 8 cores on the max while there it’s unclear if that includes the 8 performance cores or two of the efficiency ones, etc…)

  19. Could the delay in M1s because of the builds are running on a x86 emulation mode like qemu VM ? i have read at some point docker was using qemu for supporting x86 images since apple hypervisor will only provide virtualization of aarch64 vms.

  20. Finally, some real world tests for devs like me. There's just too much performance tests out there for photo/video/audio editing tools. Just proves the sad state of the M1 chips currently. Until most libraries are built/prebuilt for the arm64 architecture, the M1/Pro/Max's performance can only be fully utilized by the creatives. I would probably stick to my (hot and noisy) Intel MBP 16 until the M1 chips become more developer-friendly.

  21. Super interested comparison. Can you also do Docker Node development review? Pretty please!! For example, we are having Nuxt.js project with cca 16k of files. We are having cca 10s of hot reload inside docker container (also because of known Mac issues with docker FS.). It will be interesting to see how SSD speed and other updated affect this speed. This metrics is also quite popular among all node developers with dependencies to other containers (elasticsearch, db, api,…; currently do not have mocks for those). Anyone having some experience on this welcome to comment. Looking forward for comparison! Thanks.

  22. Comparing apple silicons with limited cores and limited RAM doesn't make sense because the cores themselves are the same for all chips. So there are no differences to expect.

  23. can you run the tests again but max out the docker allocated resources per laptop? curious to see the differences

  24. Hey Alex. That's all cool, but is RDP still slow? Just compare Mac with RDP started in Parallelised (=)) Windows

  25. can i please has some summary next time (a simple table will do the job)? and/or timeline breakdown, maybe? too many details and i can't skip because i lose context – which one you are talking about. will have to rewatch with a pencil writing down the results. because i just want to understand how these guys compare to each other. and i still don't know. thank for the effort though

  26. I am surprised why Apple still selling 8gb ram configurations. For an average developer 8gb ram is not enough. Here Apple should do the right thing and kill 8gb ram configurations.

  27. i wish more images would be optimized, i get timeout errors while running 6 containers with large wordpress sites

  28. I get my Macbook Pro 16 32gb ram 1tb tomorrow and will be running this test along with a bunch of others around Docker so I can see if I really need the 64gb version or not. Thanks for this video extremely helpful.


Install Twrp + Root Redmi Note 9/Redmi 10X 4G Stock Rom MIUI 11 & MIUI 12 (Merlin)


Como tener ya Android 12 | Novedades y Detalles