Why Apple's M1 Chip is So Fast
HTML-код
- Опубликовано: 5 окт 2024
- So we've all heard about Apple's new M1 chip, and Apple promised a lot with this chip. Including up to 3.5x faster CPU, as well as almost 4x faster video processing. All while using 25% of the power of a standard mac without the M1 chip. These are very bold claims from Apple, and the world was poised to get there hands on these chips to see how fast they really were. And now we're getting real-world results of just how fast the M1 chip is. And it's fast... real fast.
Watch this video to find out exactly why the M1 chip is so fast, and why other chipmakers are going to struggle to compete with the M1 chip macs in this sector of the market.
#TheDevDoctor #M1 #M1Chip #Apple #Mac #ARM #x86
--------------------------------------------------------------------------
Max Tech | $899 M1 Mac mini vs $2,549 iMac 5K - Ultimate Comparison:
• $899 M1 Mac mini vs $2...
Arm Vs. X86: ipsnews.net/bu...
M1 Chip: www.apple.com/...
Based on an article by Erik Engheim:
/ why-is-apples-m1-chip-...
--------------------------------------------------------------------------
Follow me at:
Twitter: / thedevdoctor
Instagram: / thedevdoctor
Reddit: / thedevdoctor
My Website:
www.thedevdoct... 
Наука
I'm a computer science student and found this video incredibly relevant for what I'm studying. This is without a doubt the best M1 video I've seen on RUclips!
Same but I’m a CE student
For real... now I understand M1 much much better.. it was mystery to previously
I agree. Best m1 video yet
AGREE!!
Computer science name is a open certificate and much people in internet doing internet works after having other certificate so anyone can inter this .
As a layman, this explanation was both eminently comprehensible and highly detailed. Truly outstanding presentation. Keep up the outstanding work!
Thanks Paul for the kind words! I’ll do my best to keep bringing you great content! Thanks for taking the time to watch.
really good!!!
@Igor Velky You are out of context and sound butt hurt that Intel is STILL behind since Broadwell. Deal with it.
As a layman, I truly agree with this assertion! Keep up the good work mate!
Great job! CISC is doom!
Steve Jobs once said: “Simple can be harder than complex: You have to work hard to get your thinking clean to make it simple. But it's worth it in the end because once you get there, you can move mountains”. I think this is what you just did, and I loved it! Thank you
Steve jobs never designed a chip and M1 is not simple.
How do you move mountains with simplicity?
@@Teluric2 my comment is about the video explanation, not about designing or making the chip
This was a wonderful explanation of the differences between the architectures.
Totally agree. It's the best explanation I've seen and one that needs to be seen by others. His critical points demonstrate how out of date AMD and Intel chip designs are compared to the M-series and why Apple's corporate decision to vertically integrate everything is the only way to achieve ultimate speed. Yes, there will be custom PCs that could be made faster than Macs but nothing off the shelf will be comparable. Apple's design is so much better and faster that they will be able to blow anything away from any other mainstream computer vendor. The first three M-series devices are entry level, yet they have the fastest single CPU Geekbench scores (not counting fake Ryzen-based iMacs which are being operated outside the legal use documents) with multi-core and GPU results being more than decent when considering how inexpensive they are. I'm sure we'll all be blown away when the next version of the M-series is released. Single core might not be much faster (could be with increased temperatures) but multicore and GPU results could more than double if not quadruple the initial results. I've seen people say a 32 core M-series could cost $30K but I disagree. The M-1 SoC probably only costs Apple at most $200 so a 32-cores might only cost them less than $1000. I may be way off but if I'm not we could see an iMac and Mac Pro really dropping in cost without hurting Apple's profit margin.
it seems like he got all the details from a blog post i read a few weeks back which is eerily very similar to how he is explaining it even the step by step process
@@delta0307 do you have a link to it. Must admit the video format is sometimes more convenient.
BTW he does link to a written article on the same topic in the description.
@ Try typing the title in google. There should be an artible/blog from dubugger.medium
For anyone else: debugger.medium.com/why-is-apples-m1-chip-so-fast-3262b158cba2 (assuming this was it).
Would love to learn more about arm and x86
if thats the case do CISC vs RISC also then
Chip instructions are sort of like pixels on a computer image- they are the most basic units that programs can be broken up into. They are also hard-coded into the physical structure of the chip. Each instruction is its own series of transistors and circuits that can complete that particular instruction. ARM and x86 are chip instruction architectures that have their own sets of instructions, and their own approaches to making instructions.
ARM and x86 belong to two different classes of instruction architectures. ARM is a RISC architecture and x86 is a CISC architecture. They are the most widely used of their kinds, but not the only examples of each. Neither are static- they develop a larger number of instructions over time. RISC architectures use instructions that are all of the same length, while CISC architectures can use instructions of a variety of lengths. Doesn't seem like such a big difference, but it means that ARM (a RISC architecture) can more efficiently manage computer resources. It took a lot longer to mature because it's fixed instruction length meant that it was difficult to implement enough instructions to do complex tasks- and it uses more memory resources. In the modern era though, those issues aren't such a big deal. It's complex enough now to do anything you'd want it to do, and memory is a lot cheaper today. The power and heat efficiency also means that you can run those chips faster. It's advantages have been especially useful in small electronics- an increasingly expanding use case- and its development has seen it used in more complex applications.
You can kind of think of the difference between the two types of architectures as a game of tetris. CISC is a regular game of tetris- all the shapes that fall down are irregular in shape and unpredictable- but with a bit of strategy, you can fit it all together and eliminate a row. The faster it goes, the harder it is to strategize quickly enough. Getting a high score is hard. RISC architectures are like playing a modified version of tetris where all the shapes are the same 2x2 block one after another. It would be a lot easier to get a high score, huh? There'd almost be no difference between a novice and a master. And the game would have to go A LOT faster before you got overwhelmed and failed.
Or at least, that's my understanding of them. I could easily have gotten something wrong.
I can see you becoming a big RUclipsr very soon.....amazing quality video.
Great explanation! I'd love a video on ARM vs x86, with details on how the M1 emulates x86 so well while others (Windows on ARM) do it so poorly. I've heard it has something to do with an extension to the M1 instruction set, but I'd love more details.
Watch this space
I was like, “This guy is good! How come he only has 2k subscribers?”
Then I checked and it’s a fairly new channel.
Looks like this would be another one of those very popular channels that I subscribed to in their early days.
Man, this video was extremely informative to me as a Software Engineer. I have studied this RISC vs CISC stuff in my computer architecture course but didn't ever understand why RISC/ARM was so power efficient as compared to CISC. Your video has made clear the advantages of locality of reference in M1 chips and the parallel decoding scheme due to its simpler instruction set. I would really appreciate if you continue to explain some core Computer Architecture and OS concepts applicable to the ARM vs x86 war. Anyways, great video on nailing the fundamentals of why M1 has outperformed x86 arch.
This is an incredible channel. All of the visuals you created to help explain something incredibly complex, possibly one of the most complex inventions in human history (a CPU chip) were so helpful. I subbed to help you get more visible.
Man.....this is the best video on M1s inner workings. Thanks for uploading.
Great video man, only 192 subs.. how is that possible. Well you've got 1 more now!
Thanks so much for the support Vincent! Not been going very long but let's hope to add a 0 to that number this time next year!
@@TheDevDoctor Well, you did within a few days 😳 ! Joining your subscribers as well, thank you !
3 more.
No one did the job of explaining better about M1, till now than uuuu👍🏽😍
Finally a different video on M1 chip. answered many of my questions
I can't believe you only have less than 1k subscribers. You deserve more. You have all the potential to make it on top.
I was not expecting such a good explanation. Thanks for actually going into the architecture of the chip with useful detail!
You just explained a 4 month Operating Systems module from University in under 14 minutes that was way more digestable. Great work!
I'm just some guy who is shopping for a new computer. No developer background, can't code anything and I was so impressed with the way you explained all of this information I was forced to come on here and comment. Kudos to you man, relaying information in such a concise and digestible manner is a show of your level of understanding. Thanks for this video.
No one explained m1 better. I was looking for the reasons and you nailed it. Subscribed.
Perfect level of details. I'll share this with my company's IT dept. Thank you for this well thought out and very professional explanation!
dude you deserve so much more subs, this is great content
Wow... One of the most in-depth explanation about the new M1 chip. Thanks a ton
You explained OOE without introducing any inaccuracies and keeping it very simple, good job :D
4 vs 8 decoders: It's not a reason for being faster, but the proper balance based on the instruction set. The RISC is already simple opcodes. Intel's instructions generate multiple micro-ops from one instruction. The term "micro ops" is basically translating the Complex Instructions into RISC instructions. E.g. ADD [RBP+8*RBX+constant],EAX means:
1) calculate the effective address
1.1 RBX shift left 3, stored in TEMP1
1.2 TEMP1 + constant, stored back in TEMP1
1.3 TEMP1 + RBP, stored back in TEMP1
2) load: TEMP2 loaded from address in TEMP1
3) add: TEMP2 + EAX, stored back in TEMP2
4) store: TEMP2 stored at address in TEMP1
That's 6 micro-ops and multiple temporary results, one of which must be held until the end, from one instruction.
The Intel chips have multiple decoders some of which are more capable and others handle only the common instructions, but it's limited usually by how fast it reads in program memory.
The *counterpoint* is that the instructions are very dense, compared to RISC, so more instruction fits in a given amount of memory; in particular, the C0 cache and the gulp size of the prefetcher and decoder.
4 decoders (with a complex + simple blend) is matched to how many instructions can actually fit in one "gulp" as it loads program memory.
The actual improvement, that affects how you would balance things, is the *improved memory bandwidth* thanks to tighter SoC integration. This means you come out ahead by having simpler "pre-digested" instructions that you don't have to work to figure out at run-time, saving the silicon budget for the complex decoders and the critical path of figuring that out before doing anything else. You trade that off by needing to read more instruction memory, faster; but that's exactly what you have.
Since the "decoding" is essentially transparent, you don't rely as much on a micro-op cache. I suspect that the ramifications of having such is rather different on ARM.
Decoders only exist with RISC based stuff right
This vid definitely deserves more views as well as you deserve more subs.
I would have thought the scope of your video to be too broad to cover effectively, at least in a reasonable amount of time, but you did an impressive job of covering some pretty esoteric topics at a level which is understandable while still being genuinely informative. Subscribed!
this was amazing. i have not seen such detailed video on M1 any where on youtube.
Great explanations. But i think nanometers also needs to be taken into account. Using 5 nm from TSMC is one of the key factor in its amazing performance.
Well done, it was about time someone presented this in a language understandable for normal technical minds
Very good in-depth review. I would add on more point, due to the complexity of the x86, it takes more die space per core. Thus, it has less room for L1/L2 cache.
Best video on youtube when it comes to explaining the performance of the M1 chip!
very well put down and explained. This is very welcomed info to understand the behind the scene of M1 chip. Thanks a lot for that.
I don't understand a lot about chip architecture but this video helped me understand the answer to actual title of the video. Keep the good quality information coming. Subscribed!
A video on x86 & ARM would be great indeed! You explain really well!
Only got this video recommended now, superb explanation! Thank you
This video made me feel proud of my purchase. Great video btw.....you deserve a lot more views.
Thanks, Sivamani, yeah any M1 purchase is a great buy my friend. Thanks for the kind words!
I randomly stumbled upon this video, popped up in my recommended. Dude, GREAT job on this breakdown/informative video. I learned so much about the underlying workings of chips from this.
I envy your ability to explain things so comprehensive yet so simple, great video!
Just the right amount of abstraction, with necessary details mentioned.
x86 has been out-of-order execution for a long time. It's only more recently been done on ARM, and not a reason for ARM instruction set being faster.
He was crediting it to more the 8 decoders compared to the 4 on x86. The out of order explanation was just him telling us what the decoders do.
I love how you explain this and yes an ARM vs. x86 video would be nice
When I first saw the thumbnail of your video I thought: great, yet another video on the M1 chip… But then I watched it, and oh boy I must say that I am impressed. All your explanations are crystal clear and seem really relevant. You got me really intrigued over the response from Intel and AMD to this. A thank you is in order!
This is what i always wanted in tech channels...thank you so much for this great explanation
Excellent presentation of highly complex concept. It has answered all my questions. Excellent!
A simple but much needed explanation of some fundamental issues, thanks!
This is the first time I understand the concept of cpu after learning at school couple years ago.
This video could be the lecture on Computer Architecture. Very well explained.
kept looking for this kinda explanation. its beyond great. please explain the difference between cisc and risc
Man! Awesome in-depth video and very well explained, though i don’t understand each everything in depth, i’m not a computer science student, but did get a lot of it. Thanks
Bruh this is the video ive wanted for a while. Most videos explaining cpus are either too dumbed down to actually get an understanding, or on the other end they are too complicated to get what they are saying
i watched the other video with the same title, but what they do only retelling Apple KeyNote...
while you really nailed it...
Very informative video. Thank you for posting
Thank you so much. This explains exactly what I’ve been wondering.
you, sir, are amazing. As an ordinary web dev, I never really understood these things and now you got me very curious about this whole domain. Please do more videos explaining computer architectures to devs
First time watching a video from you. Keep up the good work, love the in depth analysis!
want to see your channel to grow fast, nice video
Outstanding video! Really! And also please do X86 and ARM instructions set video too!!
The scope, clarity and everything about this video was superb ! Thank you
Truly the best M1 explanation I have seen yet. I shared this video w my 11 yr old because I think the video is so well laid out, that even a child could grasp. Definitely in line with the KISS acronym-“Keep It Simple Stupid”.
WELL DONE!! I have just subscribed & look forward to your future vids.
Please do a video on the difference between Arm and x86. Also please consider doing one on RISCV. Fantastic video.
excellent vid, informative but not overwhelming. u earned my subscription
Awesome explanation - Remembered me of the Advanced Computer Architecture paper !!!! Keep it up..
Im hopping on the boat! - great explanation. Easy to understand as a rookie.
This explanation was nothing short of beautiful
I learned so much watching this. Thanks for making this video and great work on it...very effective content and motion graphics!
Great explanation. By far the best video on the M1 chip I've found
Very nice explanation of the actual technical difference of the M1 that matters in real world usage. Btw, just got my first M1 machine. As a old school IT person all the way from the 90's, I think this is a piece of marvel. Oh, i'm not a coder my self, but your examples gave me some nice old fuzzy feelings of assembler code and stuff 👍
Your channel is a Amazing please keep doing Mac things related to Mac. It’s so rare to find someone who knows these things. I’d love to learn about how apple process photos and videos. And just anything that is. On the software side. Hope you get to 1 million subs ❤️
This is literally one of the best explanations one the m1 chip's advantages on the internet
This was so amazing to watch. Well done. Thanks.
Thanks for making a difficult topic easy to understand.
Surprisingly well detailed, as someone currently studying ICs i can say this guy really understand what he’s talking about, u deserve more subs, subscribed, keep it up mate.
Even though, as a CS graduate and currently working as an SRE, most of this is already covered by uni, this is still an incredible technical dive into the lower-level designs and is by far the best one on the site. Definitely earned a subscription from me.
Concise. Clearly presented. Perfectly paced. Excellent video. My first sub ever.
Excellent video! I got tired of the videos capitalising on the release. This was really informative and really well produced. Thanks
Very interesting video and very precise and detailed explanation. Thank you very much!
An excellent synopsis, thank you!
Glad you enjoyed it!
Dev-D that was well done. Thank you.
really great video. Great graphics and explanations. Please make another one on risc vs cisc
Thanks for such an in depth video, I had not considered the lack of scalability in x86 decoding before. They must either be confident they will only need 4 bytes for all their potential instructions or have a wider decoder developed or in development. I am not convinced that discrete components are dead quite yet; they do provide some advantages such as upgrading only what you need to upgrade(GPU vs CPU) and distribution of thermal loads. Ideally, you would vertically stack various components on each other as thermal limitations allow and have the entire system in an integrated block of material. I don't think we could reasonably do that these days, but that something like that could be the the future of traditional computing.
The best explanation i’ve come across thx so much subbed
Glad you enjoyed it, let me know if there are any other topics you would like me to cover 👍
I’m actually a medical doctor with an interest in tech / med tech. Also doing a computer science Masters.
@@TheDevDoctor you know more than the top high paid you tubers with many followers i congratulate you sir.
Thank you very much I really appreciate the feedback! I look forward to sharing more videos with you.
Awesome explanation Matt! Keep it up.
Thank you. To this layman your explanation was concise and coherent.
Nice good job summarizing complex comp sci concepts :)
This is very good information and it reinforced my buying decision getting the Mac m1.
Thanks for your time uploading such great content. Keep it coming.
Good explanations, good visualisations, good breadth and depth of coverage - I rarely susbscribe to new channels nowadays, but today I give all the ordained offerings to the Algorithm that it might be benevolent to your channel.
instant sub. i was mired (sort of dating me) in itanium and alpha. never had an itanium but was reading everything i could before it was released. risc-ish, but more instructions than most risc chips and intel put so much work into branch prediction. i worked at an isp back in the mid 90s and we had some DEC Alphas ;) eventually i bought a power mac just to mess around with but i have been using intel or amd at home other than an apple2c and a g3.
Very well done!! Your explanation of a very complicated topic was excellent. Can not wait to start following you and seeing your other content. Well done!
the most amazing and most useful M1 video on youtube. this is what I wanted when I watched 20 M1 videos that were maybe sent a laptop by apple IDK but none of them explained the why part. thanks mate, I will subscribe.
Amazing. This video, by far, is the most detailed analysis on M1 that I've watched till now.
Hey, thanks for the explanation. Easy to understand a complicated subject. Thanks "Dev Doctor" . Hoping you will have more of these types of videos soon
This video is phenomenal. One of the most clear descriptions of what's actually happening on apple m1 chips.
Nice to see a video talking about the technology and not another benchmark. Great video.
I can’t really add anything that’s already been said. Bravo and thanks for making this video. Going to definitely share this.
It's amazing that you have less than 2K subs and are producing videos of this quality and detail. Keep it up. I'm subbed to see how fast your channel will take off. BTW, an ARM vs X86 video should kill it in the youtube algorithm with the right keywords in the title... something like: ARM vs X86. How the Apple M1 is stomping on Intel and AMD.
Good job man! Very interesting video; well done.
Dude! Awesome content, keep it up!
There is a huge demand for a fairly detailed content with a lot of related notes and references.
Very well explained! ... subscribed.
nice explanation.... because of that I decided to subscribe... keep up the good work.....
Excellent video :) thanks very much for making it!