M1 Chip is as FAST as M1 Max!!! 🤯 (13' Mac Air Vs. 14' Mac Pro) #shorts

🛑 NOTE: Under multi-threading testing conditions the M1 Max is of course faster than the M1. BUT as a data analyst, I more frequently run single-threaded code; therefore, the benefits of paying +$1K for these minimum gains are not worth the money. I've also tested many other multi-threaded applications for data analysts and don't notice much of a performance increase. Because of this, I recommend the M1 Macbook Air for those starting in data science.

🌟 OTHER NOTE: I'm not actually selling my M1 Max... sorry. 🤣. This is a great computer if you are working with tools outside of the data science realm (like video editing in my case).

🐍 Technical details of test 👉🏼 A 100,000 iterating loop test requiring list comprehension for paragraphs processed with NLP. Test was performed multiple times on these computers to verify results.

💻 Computers:
2020 13" M1 Macbook Air - 8GBs Memory
2021 14" M1 Max Macbook Pro - 64GBs Memory

Certificates & Courses
==================================
DataCamp:

Coursera:

Build a Portfolio Website
==================================
Rebate Code: "LUKE" (Save up to 91% on annual plans at checkout)

Recommended Books
==================================

Tech for Data Science
==================================

Social Media / Contact Me
======================

As a member of the Amazon, Coursera, Hostinger, and Data Camp Affiliate Programs, I earn a commission from qualifying purchases on the links above. It costs you nothing but helps me with content creation.

#datanerd #dataanalyst #datascience

[jolness]

Python is notoriously hard to multithread though so it makes sense. For single thread of operations, they operate at almost exactly the same speed and unless you’re getting to the point where your saturating the heat sinks in the MacBook Air (8 or so minutes with all cores loaded) there will be no performance difference. On multi core aware apps, the m1 pro and max will be faster or in heavily memory bound stuff.

That said, the air is a HELL of a performer and with 16GB of memory, would last most folks 5yrs or more easily. We just got m1 max machines with 64GB (only reason we went max) of memory and it’s a huge jump vs my old 16” i9 machine.

[MHBAQ]

To test:
Normal people: write code in python
Every programmer ever: mhhhh let's steal some code from stack overflow

[BlueSquad]

They just leave a 1k computer on the doorway? America is funny bro

[prabhsaini]

the pain when i heard python, it was absolutely unbelievable

[ArtAniStokuv]

Luke you became my virtual mentor for how to be DS nerd :D great video

[timlawson]

It really is the best bang for your buck. I am a professional video editor and absolutely love my M1 air 16gb / 1tb as a secondary system for when i’m on the go. I have a very effective proxy workflow that doesn’t require all that much horsepower. The pro models are vanity purchases for 75% of buyers imho.

[himynameisryan]

I’ve had the base model since launch, it’s SUPER fast for my video editing workflow, but it can’t keep up with my 6 streams of 4K.
The m1 max is designed for video editing with extra graphics. The m1 pro and max have the same cpu

[blurrbii]

the storytelling in your video short is so good

[sheal]

Dude thank you so much before this vid I was on the fence and pulled the trigger at $800 the very same day I saw this clip, very next day m1 13in air was back to $900

[YD_.]

That’s cause the single cores are the almost the same. The main focus is the multicore because M1 Max is a scaled up version of M1. This test was based on single core and tells you almost nothing in performance between the two.

[joshparra]

Yo this is CRAZY! Snagging me a M1 Macbook Air for Christmas. Thank you for this!

[krithikramesh]

I don’t know what tests you ran so bare with me here, but isn’t the limiting factor here that python natively used single-threaded processes?

[Robstrap]

I love how all the mac YouTubers were saying the M1 was the best thing since sliced bread and that it’s incredible. Then a week later when the new macs came out suddenly it’s slow lol. Y’all can go spend over double the price to save 2 minutes on a video render but il stick with my original M1 which is more than fast enough for 99% of users. Idk why so many people are willing to spend an extra 1k to save 5 minutes on their renders they do maybe once every few days

[akshayrathore]

Run a parallelised code.
Serial code will take similar time even on a super commuting cluster

[atinjanki]

Cool. Carry out some more tests 😀 would be helpful for someone like me, who's planning to buy M1

[pepazetek]

Well, In fact you’re actually testing the single core performance. Since the M1, M1 Pro, M1 Max and M1 Ultra are all built using the same technology, they also got same single core performance. The differences are in the number of CPU and GPU cores. Sure, this test case may show many people’s use case, but I am just saying that if you really wanted to get the best use of the potential of any chip with 2 and more cores, you would definitely implement multithreading. In that case, you really see the difference (approximately the same as shown at geekbench websites).

[Mostlyharmless]

Cool cool cool. What’s the difference for swift though? Because that’s what matters for macos

[natlamy]

At first i thought this test was stupid, but then i thought of what mac users do, witch is basically only desktop use, so i realised that its actually a good comparison of performance.

[MangoTeee]

The reason why there’s such a small amount of increase in performance is that python is extremely interpreter limited, meaning that past a certain point most computers nowadays won’t run python much faster than an expensive PC (unless the python program is programmed for multi-threading, which doesn’t happen because python is kinda a joke language if your tryna optimize code)

[riley]

Single threaded test on computer with the same cores but more of them. Very nice