Apple M1 Pro & M1 Max Initial Analysis - Desktop Zen 3 CPU and RTX 3080 Performance?!

Patreon: / fullybuffered
Apple has announced two exciting and enormous new chips for their new Macbook Pros; the M1 Pro and M1 Max along with an impressive list of performance charts. Today we take a close look to see what kind of performance that could mean compared to desktop parts.

[@PiXimperfect]

Woah! Now that's a video I needed to watch! You, sir, are brilliant! The best analysis I have seen on KZbin. Subscribed!

[@FullyBuffered]

Thank you!

[@Changbinsww]

Now this is a really informative video! Thanks for the hard work.

[@FullyBuffered]

Thanks!

[@moosicman101]

Finally, something useful that’s not “things Apple DIDNT say”

[@FullyBuffered]

Thank you!

[@lordvader4255]

Really sick of those videos that is just a clickbait rather than real data

[@paulmontgomery4696]

You earned another subscriber pal with your articulate presentation of analysis of the data and information your research efforts acquired, gleaned from the assertions in Apple presentation, and referencing your sources throughout the video. Looking forward to viewing the videos yet to come from this channel and may many more subscribers follow. Job well done.

[@DavidHarry]

Excellent video 👍 Cheers, Dave.

[@FullyBuffered]

Thanks Dave!

[@andrewcox9230]

Thank you so much!!! Great video. Very very helpful.

[@FullyBuffered]

Thank you!

[@ArcticPathway]

Had to sign in to subscribe because I see a bright future for you in this industry. Thanks for this informative video.

[@FullyBuffered]

Thanks!

[@athreyaabhijith]

A sub for sure for the relaxed commentary and detailed explanation.

[@FullyBuffered]

Thank you!

[@VRcation]

I typically only subscribe to someone after seeing 2-3 videos of theirs that I like, but I subbed at 0:09 upon never seeing this channel.

[@FullyBuffered]

Thanks :)

[@francisyuweh706]

Looking forward for more analyzing and comparison w/ these new m1 chips. Keep it up!

[@FullyBuffered]

Thanks!

[@MrHurdleking24]

Great explanation abs comparison! Subbed

[@FullyBuffered]

Thank you!

[@ilfaraone58]

I like your review style! You deserve a sub for sure! Keep up the good work

[@FullyBuffered]

Thank you! :)

[@patrickgaerlan]

Thank you for this 👍🏻

[@MaraAmaraaaa]

this is so good -- you just got yourself a sub

[@uesugikenshin8430]

I think you do the math wrong in the 10:59 chart. It shall be 12550/400*375 = 11765, whic is exactly the same as your Razer estimation of 11872

[@voltazkai]

Yea that’s what stood out to me too while watching the video. M1Max performance should be lower than than the GE76 Raider. So the second calculation should only help reinforce the performance claim of the Razer…meaning it should be very similar to one another. So thank you for the calculation. Nonetheless, still have to give credit to this video. Because it’s one of the few videos that actually tries to make sense of Apple performance claim instead of parroting whatever Apple said in their presentation.

[@FullyBuffered]

Oh shoot, nice catch, you're totally right! Then it lines up even better indeed!

[@FullyBuffered]

@@voltazkai Thanks!

[@jmsobe]

Excellent analysis, you earned an sub. Make you wonder what they will do with the Pro Desktop machine when they unleash the wattage without battery constraints.

[@FullyBuffered]

Thanks!

[@imjooboy]

Did you take into account what effect the increased memory bandwidth will have on different workloads for the M1 Pro and M1 Max?

[@b_markovic]

Apple's charts already account for that.

[@imjooboy]

@@b_markovic WIthout knowing the specific benchmarks they ran, it doesn't necessarily account for it on the CPU side. Many cpu benchmarks run fully within the cache of the CPU and don't go out to main memory very much.

[@b_markovic]

@@imjooboy Well how would you imagine him accounting for that. No one could account for that without running real-world apps really, or benchmarks specifically aimed at a mixture of FLOPS and RAM throughput / frequent heap access etc.

[@imjooboy]

@@b_markovic Its clear you aren't familiar with CPU benchmarks and different types of compute workloads. And thats OK. It might be that this guy is familiar with them and it might be that he's not. If he isn't, thats ok too. But I won't find out unless I ask. And on the off chance he is, that should be an area of the M1 Max performance that should be performed because it's high bandwidth memory is probably great for certain types of workloads.

[@b_markovic]

The only thing clear here is that you haven't watched the video or you have absolutely no idea what you are talking about. He extrapolated the figures FROM THE GRAPHS PROVIDED BY APPLE. Again, HOW WAS HE SUPPOSED TO ACCOUNT FOR ANY PARAMETER WHEN HE IS EXTRAPOLATING FIGURES FROM A GRAPH?

[@blakerupert]

Great video. editing is snappy, delivery is smooth, calm and informative. How'd you get the red laser for a cursor? Are you using OBS and Browser view?

[@FullyBuffered]

Many thanks :) I used MS PowerPoint & Nvidia Shadowplay for this

[@jeffyeung6103]

Your content is pretty interesting. I recommend having the mic closer or speaking louder. It's kinda hard to hear. Anyways good job! Really tempted to get the M1 Max 24 core

[@FullyBuffered]

Thank you!

[@troder5428]

Really liked your video..

[@FullyBuffered]

Glad to hear!

[@oscarmvl]

I want to be skeptic of these charts, but at the same time, the M1 destroyed the competition while providing excellent battery life. I guess we’re just a few days away from finding out if the hype was real.

[@RunForPeace-hk1cu]

The m1 hype turned out to be real while LTT made fun of apple. This time around, LTT will be much more reserved so that don’t appear looking like complete fools. The hype is real

[@jmpattillo]

This makes me wonder how insane the Apple Silicon Mac Pros are going to be. They will be without the need for power constraints and command far higher prices.

[@ppietrap3876]

let us put it this way, the Mac Pro will have an option to install a SoC with 40 CPU cores and 128 GPU cores. So we can imagine at the very least 3 times the performance when compared with the M1 Max.

[@jandersen6802]

Perhaps this huge computing power will attract a lot of gamers and game developers in the coming years.

[@12pure]

@@ppietrap3876 dont expect more than 32 performance cores and 4 efficiency cores

[@PTYC933]

@@ppietrap3876 likely close to 4x. @juice fruit, the jade 2 c is expected to be essentially 4x the m1 max. Maybe 32 high 8 efficient.

[@b_markovic]

@@jandersen6802 Those prices are prohibitive for people that don't actually earn money with that computer -- which excludes 99% gamers, and game developers won't really stay in the business by targeting 1% of the market.

[@nickpunch9661]

Great analysis! Having said that all those numbers are not really useful though if we are running applications that dont run on ARM. Im running CAD applications like Siemens NX and Solidworks which can only be run through Parallels 17. So unless applications and games are native those numbers are a moot point at the moment.

[@behzad3]

right to the point also unintentional asmr subbed fast!

[@FullyBuffered]

Thanks!

[@vividmotion66]

Interesting specs. So I could roughly expect the 24core gpu to be around the performance of a rtx 3050 to to 3060? Thats be crazy.

[@evacody1249]

No you are still getting a 1070 gpu. It's no where near a desktop GPU and never will be.

[@xaratemplate]

I think you have a perfect formula for making videos keep using the same music and presentation structure and you will find yourself at 1 million subscribers. I can envision you having a very clean studio setup with very clean room with white walls plus the amazing background music used in this video. Just wanted to say great job and encourage you to push harder to keep up the good work. Never give up and keep a positive mind and you will do more than you could imagine. You might not be able to afford it the new laptop but maybe if 3,000 people donated you $1 you could buy one. Heck I would donate $1 first to help you buy one since the in-depth and honesty in this video is worthy of my support. I would be happy to see you review one of the macbooks and give an honest review.

[@FullyBuffered]

Many thanks for the kind words! I am very glad you like it and I fully intent on producing more content like this in the future!

[@FerCL1]

I came here because of the M1 Pro and M1 Max hype, however I stay because the information and the explanation is top notch. Also he’s really hot

[@FullyBuffered]

Ha thanks :)

[@ko6.]

Great job ! We will see soon if you’re right :-)

[@FullyBuffered]

Thank you! Yes so am I :)

[@jwkelley]

Is there a reason to use r23 instead of geekbench? My guess is those scores are geek bench since they show m1 is the best light. Same with the gpu scores. Even with the graphics scores with geekbench scores you can make the original m1 look like a 1650 even though if you take all the benchmark togethers its not even close.

[@sivamanipatnala5517]

GFXbench shows similar performance of 32c gpu with 3080 mobile, but in geekbench, 32c is only getting half the score of 3080...

[@jiajianhou426]

According to this, it seems my newly ordered 32-core gpu 64gb ram mbp’s performance is about the same as my 3800X + 2080S + 64gb PC. How disappointing I thought it would smoke my pc.

[@NN-br2xh]

I wanna see what the graph looks like for the M1 chips beyond the low wattage they're set at. I get that their value is that they don't use a lot of power but for situations where power is not a concern and I want pure performance, can the M1 chips do better? Or is it a case where the higher the power consumption, the lower the performance will be.

[@FullyBuffered]

Supposedly there will be a high performance mode for the 16" Max models - curious if that can push it past the numbers they've put out here.

[@maman89]

Frankly speaking x86 cpu days are numbered. I’m already begun collecting and building a few gaming pcs relevant to my era. I’ll sure miss build custom loops.

[@samirtravel1242]

Apple states that I used less power and get less performance. If you were to extrapolate the power used to see the performance numbers it could theoretically generate how much would the performance end up being?

[@Aaronage1]

Nice estimations 😎 Good to see a PC enthusiast with an open mind regarding Apple Silicon 😊

[@Aaronage1]

@Mr. Sophistication A lot of PC hardware enthusiasts can’t see the good in Apple Silicon simply because it’s Apple. They see it as a threat rather than something that will push the industry forward 🤷‍♂️

[@Aaronage1]

@Mr. Sophistication You just said “I wouldn’t use a fucking mac even if it was 10 times faster” 🤷‍♂️

[@tablettablete186]

@@Aaronage1 I think what he is trying to say is: I wouldn't use a MAC because it doesn't suppprt the software I use, so its speed doesn't matter.

[@Aaronage1]

@@tablettablete186 Yeah, don’t get me wrong, I don’t want to put every PC enthusiast in the same bucket. It’s fair to say there’s hostility towards Apple in the PC enthusiast community, though. Apple is seen as the antithesis of what they like about PC, so they’re attacked/ridiculed more often than not. There’s big personalities like Linus from LTT making silly hit pieces on Apple all the time, stoking that fire 🔥 It’s fine in general. No one has to “like” Apple obviously. I guess it just disappoints me that the prevailing mood in that community means they miss out on really impressive innovations like Apple Silicon. I would rather fellow PC enthusiasts were asking why the PC doesn’t have something like Apple Silicon already and push for that. I’d love to see a monster Nvidia Arm SoC for PC, for example 😍

[@RunForPeace-hk1cu]

@Mr. Sophistication sounds like your are butthurt Innovation is innovation. Don’t let facts hurt your feelings. If you are a true computer enthusiast, you should appreciate good computer engineering designs and breakthrough performance per watt. Human civilization is based upon building on other peoples ideas and successes. It’s called human progress. You on the other hand, is a partisan hack who pretends to be righteous when you should be looking at how other companies in the industry’s business practices are all the same. Apple isn’t a evil doer. Enjoy nut hugging nvidia and amd. Gotta stroke that ego and your feelings If you don’t appreciate 400GB/s 32channel lddr5-6400 memory on a mobile chip running at reasonably normal thermals, you aren’t a technology lover.

[@eladreltuc]

I ordered the max with 64g ram. I don’t really care either because work paid for it.

[@bevans868]

Very insightful video! From what I have read elsewhere Apple Silicon performance doesn’t change when unplugged and running on batteries but other Intel/AMD laptops immediately throttle CPU and GPU to save battery life and reduce heating. Therefore ranking should be different. Would you agree?

[@manaspradhan8041]

My ryzen laptop never throttles, it just depends on the specific laptop how it handles power delivery and thermals

[@bevans868]

@@manaspradhan8041 you are correct but some Intel/amd laptops do automatically throttle.

[@manaspradhan8041]

@@bevans868 hmmm, I wonder what I said in the second half of my single sentence reply.

[@dragosg23]

you are smart

[@LaurentiusTriarius]

Reminds me of the PowerPC days, it started so well with a lot of promises and it ended up being a two decade long mess. Idk I have a déjà Vu feel and this time Steve Jobs ain't coming back...

[@markjones8121]

In terms of performance vs cost, are the new MBP’s good value for money?

[@thinukawijerathne]

yep

[@Px4164]

only in laptop realms, you can have same power for half a price within desktop PC range

[@markjones8121]

@@Px4164 But Final Cut Pro doesn’t run on Windows …

[@thinukawijerathne]

@@Px4164 but not premium. You can't even buy a good display for that price!

[@Px4164]

@@markjones8121 and 3DStudio Max doesn't work on Mac

[@stu9000]

Enjoyably nerdy.

[@FullyBuffered]

Haha thanks that's one of the best comments I've received so far xD

[@xrypro4080]

If Apple can make M1 Max 32C better than desktop 3060Ti, I'll trade my Razer blade 15 with mac in an instance. Such a massive jump in performance (more efficient with 400gb/s bandwidth!) for the same price. btw, I think your chart when comparing M1 Max to MSI GE76 is wrong. M1 Max should be around ~11800, which is tally to RazerBlade15 estimation.

[@evacody1249]

Desktop gpu is not the same as a laptop gpu.

[@vividmotion66]

Also how do you see actual potential to play games with it?

[@ProgressiveConservative]

Not now, very few games can be played on arm, but now that they are going to be powerful enough to play games maybe companies will start releasing their games on mac, if that happens I’ll never buy a windows laptop again lol

[@cemsengul16]

The problem is not strength actually. Developers will not code a whole game two times. They make one game in X86 for consoles, arcades, and pc. They are not going to go through the hassle of coding the game for Apple silicon.

[@RH-jj7ck]

apple should really support vulkan and not try to push that metal thingy. gaming studios simply cannot afford to job 2 times especially when there isn't much of customer base(yet) for arm. I hope MS will do proper arm support. Also IF nvidia gets permission to buy ARM company(i think its like 50/50chance to go either way)... that would most likely change things quite a bit but that will take quite a bit of time(3 years i bet)

[@PixelPipes]

My graphics card is a 3060 Ti, and imagining an ultra-thin laptop with an SOC that can match it....kind of breaks my brain. I mean the 3060 Ti is only a year old.

[@adamreid5901]

3060Ti is still using old 10nm Samsung vs 5nm TSMC so I suppose it could be true

[@snowmean1]

​@@adamreid5901 first of all 3060Ti is using Samsung 8nm not 10nm, secondly decreasing transistors size can help with reducing heat and increase the clock speed, but don't actually make GPU cores faster, just cooler, hence to compete with 3060Ti, M1 should have either significantly higher clock speed, or much more efficient cores which is unlikely considering nVidia's expertise in GPU production. Another question is memory bandwidth Apple claimed: LPDDR5 has a bandwidth of 6400Mb/s, so even in 4-channel mode it will be 25Gb/s, however Apple claims 200Gbps and even 400Gbps, however I don't get how exactly this speed achieved with LPDDR5, because those speeds are only achievable on HBM2 or multi-channel GDDR6. For example RTX3070 with GDDR6 has 440Gb/s memory bandwidth. Modern integrated GPUs are limited by the slow DDR4/5 memory, that's why AMD stopped increasing GPU cores in their APUs as that doesn't make sense because most of the time those cores will be waiting for data to come from the slow DDR memory. DDR5 should help a bit, but still far slower than modern GDDR6, however Apple claims that their GPU using slow DDR5 memory will be faster than 3060Ti? Very doubtfully.

[@adamreid5901]

@@snowmean1 Nvidia's "Samsung 8N" process is 10nm+

[@snowmean1]

@@adamreid5901 yeah, and that's exactly the reason why it calls 8N.

[@adamreid5901]

@@snowmean1 Yeah so you didn't need to 'well actually' me that entire paragraph.

[@LittleRadicalThinker]

The power consumption is insane…

[@zonyyu8412]

That’s precisely the difference between designing your architectures around an iPhone and scaling up vs designing the architecture around a desktop and scaling down.

[@Harrysound]

I think part of the hype and excitement is because we haven’t had new pc architecture for years and years so no one knows where the limits of this are. I can really see intel are finished though I suspect apples pricing will force many many people to look away.

[@markarmstrong2746]

I would like to see Apple release a PCIE and MXM version of their graphics baked into the M1 silicon. That would be cool, especially if it had Windows 10 / Windows 11 drivers.

[@markarmstrong2746]

@Mr. Sophistication That is assuming that Apple will listen to me.

[@adamreid5901]

Unified memory mate

[@ppietrap3876]

@Mr. Sophistication In this case it kind of would stop Apple from easily porting the same level of performance because the unified memory is built on top of Apple’s memory controllers, plus the shared System Cache between CPU and GPU on the SoC, and the RAM being on the chip package.

[@RunForPeace-hk1cu]

Your mind is still in the Stone Age cpu/gpu as separate unit mindset. You hahe been brainwashed by the PC industry

[@RunForPeace-hk1cu]

@Mr. Sophistication they aren’t … they still copy back and forth between the cpu and gpu UMA isn’t what apple did …. Go read up

[@salman13]

isn't the advantage because of the chip size like the 5nm thing they talk about? if it's only that....eventually once intel and amd do the same...will apple really be that much better? just curious. hardware ins't really my cup of tea. thanks

[@Lewis64]

You are way underestimating Apple

[@iUmarFarooqi]

What use is all this power? You can't even play games on it 😅 I'll probably get a Windows laptop so I can play some games and also do my work.

[@Endowug]

There are people in this universe who are not interested in playing. At all.

[@cemsengul16]

Exactly bro. I just got the 11900h 3080 OLED Razer Blade and I couldn't be happier. I couldn't give a shit about Apple's theoretical power.

[@iUmarFarooqi]

I actually own the M1 MacBook Air 😅 I can't play anything on it, so I'll get a Windows laptop next time and maybe keep this one as a portable work machine.

[@known3617]

Don’t spend to much time with the graphs as they tell us really nothing. Graphs are purely marketing material for the peanut gallery for instant gratification. That being said benchmarks are coming out showing the M1 max is surprising the RTX 2080 super and 6700XT.

[@sergioyichiong7269]

The real test is running any popular power hungry app and make a video how fast the progres bar runs like this kzbin.info/www/bejne/gGLYfZ-cl8Z9rJY

[@movieblues4614]

You are not going to be able to get your hands on the Macbook Pros? Apple has a no-questions-asked 14 day try-before-you-buy period. You can do all the test you want ,then send it back if you want to.

[@zerofox3d]

Nice extrapolation chap.

[@EngJorgeSantana]

Max Pro could have 40 cpu cores and 128 GPU cores to compete with threadripper.

[@sergioyichiong7269]

NO, TR has 64 cpu cores. cards used with TR has thousand of GPU cores

[@RadialSeeker113]

Gotta work on your mic uqality

[@yusufdomun303]

What's your name brother ??

[@suit1337]

I bit sad that you have not taken into account that Power Draw increases exponetially - for example the 3080 Mobile at 100 Watts and above gets 0,8 % more Performance for every 3,5 % of power you put in. The numbers that Apple puts out are obviously real, they would not fake them - but they are cherrypicked too, like they are for every other system - remember Nvidias Promise with Ampere: 1,7x Performance (3080 vs 2080 Ti) for less then half the price - that makes it almost 4 times the performance per Dollar. Boy do i pitty those, who sold theri 2080Ti Cards for 400 Dollars after the annoucement just to realize, that in Real World Benchmarks the 3080 is mostly not really faster _and_ the street prices are way higher than you would have expected but sadly this will be forgotten and real world benchmarks have no use - only the annoucement will be remembered

[@RunForPeace-hk1cu]

Everyone laughed at the graphs apple posted about the M1 until they realized that their claims were … REAL And many cases exceeded their claims . So take that as their track record

[@zt9233]

Wait so until You hit 15w the m1 is actually more efficient and has more perform than the pro/ max.

[@Alex-zr2rr]

if you want to play gta v on parallels you get only 35fps very powerful ))))))

[@parthchopra315]

Your comparing it with Gaming Intented GPU 🙄, It's good that M1 does same on benchmark but it can't do shit in gaming, no use of it rather than editing and KZbin

[@NiiAnikin]

It's still an arm processor... No serious gaming yet

[@cemsengul16]

Yeah it will not be able to game like a Razer. Developers are not going to code the whole game two times for normal X86 and Apple Silicon.

[@ak1s4]

A ryzen 9 5980hs with a tdp of 35w has a cinebench score of ~11000 according to techspot. So with 5nm process and 2 more cores we get 2k more points on cinebench when compared to a *laptop* ryzen chip. While impressive, its nothing new.

[@suit1337]

​@Muhammad Ehtasam if you take into account that ARM is insanely more efficient than x86 BUT that only accounts for a few % of the total power consumption you get why the Apple 10 Core is power hungry as hell: most of the power consumption goes into the caches (which are usually SRAM no matter if you build ARM or x86 and of course Uncore-Power like I/O oder Memory Controller etc. Thats is why a Mobile Ryzen is much more efficient than a Desktop Part: the I/O-Die is on an older Node and the Desktop-Parts have a vastly bigger Memory Controller and I/O-Die

[@movieblues4614]

"nothing new" LOL. The only comment I had never heard before is 'nothing new'.