Is Android faster on the M1 Max than on M1 Pro MacBook?

all right so many of you requested and i ran a poll that i go and get another 16-inch macbook pro this is the new 16-inch macbook pro 2021 the notch edition with the m1 pro chip now one of these is the pro chip and the other one is the max chip which one is which i actually just swapped them out and i um left them outside to cool for a little bit so i don't remember which one is which and when i run this test today i'm gonna do it blind i'm not gonna see which machine is which and then hopefully the test results will reveal to me which one is which and then we'll check at the end so let's pop these things open and take a look all right so i've got both these machines open now and they look pretty close to identical except for the slight temperature variation which uh i'm not sure why that's happening but the one on the left is 41 degrees and the one on the right is 46 degrees other than that they look pretty much the same so what are we doing today we're doing the android benchmark uh let's uh open that up and i'll show you what i'm talking about here and you've seen me do this on this channel before but maybe some of you are not familiar and a gentleman by the name of sergey rathevsky he created this as a fork of mozilla mobile which is basically firefox for android so it's a snapshot of that repository in time so that people can run this and report their results for this benchmark using different machines and he compiled the results here so you can go to the repo check out the results of other people and their tests there's a google sheet there but this also tells you how to run the project what are the requirements and i've set all this up i've also uh posted a video how i set up android studio from scratch on an apple silicon machine which is not exactly straightforward anyway let's uh get beyond all that and we're gonna go and pop open android studio on both of these machines and there we go i have android studio benchmark on both of these let's open that up now by the way you don't need to open android studio to run this you can just run this from the command line from the terminal but i wanted to have android studio open as well because well that's how typically we use android studio not through the command line but most of the time we use it with the android studio open people that build across platform apps don't have to do that the clies usually handle that for you so you don't have to but we're going to today so here we are at the terminal down at the bottom and i think i might have given this away already at least to myself i can tell the difference between those two terminals because i know what i called the machines sorry folks busted oh well everything else is pretty much equal so let's go ahead and run this for those of you they can figure out good for you for those of you that can't maybe you'll be surprised at the end heck maybe we'll all be surprised at the end so to run this uh you use the command gradle w clean assemble debug all right we are ready to go and i'm going to hit enter at the same time because this time we don't have any dependencies to download so why not let's make it a race those are always fun and ready let's go okay the one on the right seems to be faster already at least at that part let's see what happens here yeah uh the machine on the right seems to be leading because this one is not even moving maybe it's doing something i don't know but this one well i can't really tell to be honest with you by the output of this uh build so we'll be right back after this is done we'll review the results oh and just in case you're wondering the specs for these machines will be in the description but they are both 10 core cpus okay we have a winner but not by much the machine on the left actually has one minute 48 seconds and the machine on the right one minute 53 seconds so we're not gonna go by just one number we want to know the average so i'm going to do this a couple more times and let's go if we get consistent results where the machine on the left is constantly winning i'm gonna go ahead and say the machine on the left is the m1 max and just a little review of the current temperature situation this is 63 degrees right now this one is at 54. let's take a look at the battery here on both of these machines oh batteries at a hundred okay so they're both exactly the same right now we've kicked things off they were fully charged at a hundred percent so let's see how the battery gets affected after a few builds okay this one is done at 125. this one is still working at it wow 152 pretty consistent with the machine on the right but this one just uh took off and again this battery is really holding strong isn't it they're both grabbing on to that 100 battery power i think that after this test i'm gonna run one more test and then i'm gonna play i don't know some videos to see which one of these will be first to 99 battery that's not a long-term use scenario of course so don't judge a battery by the first percent lost but that's going to be for a separate test the battery test will have to be probably a separate video let me know if you want to see that and if you do want to see that let me know what programs you want to see drain the battery now as far as the temperature of the cpus they're hovering around the same area around about 60 degrees 58 to 61. so i don't think one is heating up more than the other one so i'm not really getting an indication of let's say a really powerful hot gpu or cpu well gpu's not really being used here but the cpu right however since they're both 10 cores maybe we won't see any difference and here we go look at this now we've got 123 on the machine on the left and we got 125 on the machine on the right i don't know i don't know i think we need to run this again that was so close that as soon as i looked away from one the other one was done folks uh i don't know if this is making much of a difference right now let me just write all this down i mean clearly the one on the left is winning consistently but but the margin is getting smaller and smaller each iteration of my runs here so if i had to guess i'd say the one on the left is the m1 max now what does cause a huge difference in build time is which jdk version you're using are using an intel version which has to get translated through rosetta or are you using an arm dedicated version which doesn't have to be translated and i made a video on that right over here so check that out now let's check the machines to see what they actually are the one on the left here the one that consistently won is in fact an m1 max chip and the one on the right obviously is the m1 pro chip so the average build time on the m1 max was one minute 23 seconds and the average build time on this one on the m1 pro is one minute 34 seconds so a difference of about 11 seconds on average and if you're doing this a lot it might add up but it might not even be worth the extra 1000 bucks that i paid for this machine now here i've also got that 14-inch macbook pro with the m1 pro chip that i was talking about and my good old trusty m1 air this machine has really been kicking some butt the m1 air if you want to see this exact test uh racing these two machines then leave a comment down below as well now that we've got some preliminary numbers on this let's see how the battery was affected on these two machines well we're still at a hundred percent on each one of these all right so what i'm gonna do is um something unconventional and something you probably wouldn't do but this is something else programmers would do i don't know if you have a problem and you want a solution you program it right so how do we execute this a number of times over and over and over again until the battery drains a little bit and then we can check it in a little while um well use a loop right so here i've typed up a little shell script right on the shell here 4i in sequence two do that command that we were executing and then done let's change that number to something a little higher so it takes about a minute and a third to execute one of these let's do 30 iterations and that will give me time to go walk my dog and then i'll be back to check on the battery status here okay i've got that command set up on both of these let's go axel you want to go for a walk come on come on oh yeah oh yeah yeah you want to go for a walk don't you oh yeah we're a good dog yeah smile for the camera okay this thing just finished i'm back from the walk it was cold outside the the times that we're getting up pretty consistent at this point one minute 21 seconds one minute 24 seconds let's check the battery 79 80 percent so the max is using just a little bit more energy it looks like and that's to be expected so that's just a little battery test let me know down in the comments below if you want to see a more extensive battery test where we go all the way and kill the machines and i can include my intel machine as well as the macbook air and the other m1 pro that has eight cores so what's the takeaway here the max is a little bit faster than the pro if you're going to be doing these kinds of large builds now i did do a large xcode build and i did that using the m1 max 16 inch versus the m1 pro 14 inch so there's a little bit of a difference in size there and the previous pro that i used for this was eight core machine not a 10 core machine so i saw a significant difference there but this difference is not significant at all and that's probably because there's 10 cores here and 10 cores there so we're very close what's accounting for the difference most likely it's the bandwidth of the memory the m1 max has a lot more memory bandwidth than the m1 pro and that could result in these kinds of differences in a few seconds of build time if you like this video i'd appreciate a thumbs up subscribe if you're not already make that button gray and i'll see you next time thanks

Channel: Alex Ziskind

Views: 9,710

Keywords: apple m1, apple m1 chip, apple silicon, apple silicon chip, apple silicon mac, apple silicon macbook, apple silicon macbook air, apple silicon macbook pro, apple silicon macs, m1 chip, mac with m1 chip, macbook air, macbook air 2020, macbook air m1, macbook pro 2020, macbook pro m1, new macbook air, new macbook pro, alex ziskind, m1 android studio, m1 android emulator, m1 android development, m1 android, m1 android studio performance, android studio, m1 pro, m1 max

Id: BY-O2i365BQ

Length: 10min 0sec (600 seconds)

Published: Mon Nov 15 2021