😂

IIRC something similar happens with Microsoft and people like MJF - because Microsoft is so big and secretive you often don't know what other parts of the company is working on

They're not THAT smart

@@UnusedChar They're way smarter.

@@UnusedChar Smart enough to trash Snapdragon chips.

True dat. it's been really too long already!

yea

yeah that was holding back a lot of performance. long overdue.

Imo they should have included out of order execution in A55 successor. Multi core performance would be improved drastically.

@@afif4738 And probably efficiency! Maybe we can get more standby time and SoT in easy tasks with the next one.

True

amen..just let developer optimize the chipset

am pretty sure about that since apple has long time supported Macs like the 1st MacBook Pro 13" Retina was supported for over 9 years almost

@@bennaambo2716 plus Apple has supported its own processors for a very long time. For example the SE and 6s series are still supported with the latest update. And the older software versions still receive security updates from time to time

I am sure the os and Apple apps will support for the normal lifespan, but I worry that 3rd party app support could suffer. Phones will go to V9 for security and high power apps will go V9 for performance gains -but do they want to support V8 M1 for 100M market? M1s are fantastic. Better than 90% of PCs at tiny power draw, but in a few months MX MBPs will beat everything, for a few dollars more.

No way.

Why would people "who are not into this stuff" watch this sort of this content?

@@andrei.agulescu at some point you were new to this stuff too

@@Nick-jt1if Yes but I was into it. There's a difference.

*You *it's (contraction of "it is/has") *a lot (Two words, not one. Think of it like this: a few, a little, a bunch, a _whole_ bunch, a lot, a _whole_ lot)

It will come. ARM and Risc-V is the future. This feels like the mid-90's all over again. Back then we had talks about how Risc would be the future, and not Cisc. Better late than never. 😉 Our dreams from back then, might come to life in the next years. Byebye Cisc/x86 and hello Risc.

*plunge (different meaning and pronunciation)

@@alvallac2171 absolutely right, thank you.

Yes at last God has spoken

Introducing the new A57... wait

Man I knew I couldn't be the only ones sick of a55 cores lol

Fuck A55, shit too weak for 2021

@KNAX_NO4H how is that so?

SystemReady isn't a Armv9 thing. It already exists for Armv8.

Competitive is relative. Depends on what light you look at it under. Raw muscle, then yes. Power consumption and heat, perhaps not so much.

It is not like that. ARMv8A, v9A etc are instruction sets only. Software developers use the instructions to write a codec. Some vector instructions can help to speed up the codec if those instructions are being used. Rarely do they put a hardware codec inside a processor. Google does develop custom 'Argos' video-transcoding chips for UKposts processing. It supports VP9 currently, no AV1 yet.

@@catchnkill thanks for the clarification dude.

There is a scheduler in the OS. Here is a good starting point: www.androidauthority.com/arm-bringing-energy-aware-scheduling-to-android-664926/

*cores (plural) core's = possessive (or a contraction of "it is/has") Apostrophes are for contractions or possessive nouns, not for pluralizing. *other. Is (to fix your comma splice run on) *piece

It isn't about performance, it is about features. The architecture like Armv9 doesn't define or set the performance at all. That is the micro architecture (the actual chips design). For example the Cortex-A55 and the Cortex-A78 both use the same Armv8 architecture, but their performance is (intentionally) very different.

exynos will forever be shit even with amd gpu

@@xtian9146 lol, isheep spotted

Exynos 2200 might not have arm v9, because it's launching next month

@@DrAdityaReddy sheeproid spotted

Sorry, no.

@@GaryExplains why ? Then please do something similar to it or best make the last video on how to write your own architecture with actual source code

Thank you for remembering this too. So many ideas - like virtual machines and clusters - have been in use for a very long time. HP-UX forever!

Also - OpenVZ - different beast, but still...

No. If Qualcomm is jumping to Armv9 this year it will be announced in December.

SD plus SoCs are usually the same silicon as non-plus, just binned differently.

Name a victim of spectre

Yes, there was some info about Mali, specifically that it would support Ray Tracing and Variable Rate Shading in the future.

I think it makes perfect sense. Apple typically gets a 15-20% boost on a normal year. Around 30% boost moving to a new nm process. So if they use both they will be stomping the desktop & laptop market 👀

@@chidorirasenganz The chip architecture is completely independent of the fabrication process. Yes, performance gains can be had from doing either shrinking the die or moving to a more performant architecture (or both), but if the fabrication partner (TSMC, in this case) can perform a die shrink, that's far easier for Apple to plan for than switching to a new architecture. That's why I'm skeptical we could see it so soon. That said, Apple's got some of the best engineers in the world working for them now, so maybe they can. I guess we'll find out.

@@RobStevens64 Thats true and exactly what I was getting at. They aren’t mutually exclusive. It was recently announced that Apple was going to 4nm for the M2 Mac. So I could see apple doing both to solidify their lead. Apple went 64 bit before anyone else, I dont see why they wouldnt adopt ARMv9 unless they would rather have a modest boost in performance

Yes

I hope so

Hope we see M1 like processors and Windows 10 built from scratch on linux kernel to match performance of macs. It's disheartening to see how slow Windows machines are compared to Macs...

@@rupeshvislawath windows machines aren't slow. Lol

Yes, 100% compatible at a machine code level, however the OS and hardware would be very different. I once compiled a small program on and for a 32-bit ARMv6 raspberry pi and ran it on a 64-bit Armv8 Snapdragon, worked perfectly.

@@GaryExplains that's amazing, thanks

Performance comes from the microarchitecture, not the ISA. Think about it, the Cortex-A55 and the Cortex-X1 both have the same ARMv8 arch.

There’s no valid point to that: do computations as to how large a 64-bit address space is, keep in context how memory is read and written a cache line to/from main memory at a time, and CPUs already handle the larger data chunks at a time for media mathematics, and general-purpose data handling doesn’t benefit from more than 64-bit (ok, IEEE doubles are 80 bits in size, but again, see the rest) combined with fixed width instruction and operand layouts for RISC makes decoding so much easier you can scale OoO better, adding the instructions to support 128-bit instructions would require more memory bandwidth and power used for processing: it’d be very wasteful and would actually slow down the system. I’m currently 49. Perhaps before I die, people may have 64-bits worth of actual RAM on machines, since I remember first using TRS-80s with 8KB RAM, I now have a laptop with 64GB RAM. So, in about 39 years a fairly readily available computer (but not the cheapest for the time) has gone from 13 bits of used address space to (and I’ve written code for a machine at work with 1 TB RAM, 120 logical cores) to 36 bits of physical address space filled with actual RAM, so 25 bits. Oh, and my laptop will readily run over 7 hours on battery in most things, until I start compiling a bit of code or play Cities:Skylines :p We’re at a point for hardware that whatever can theoretically be done in the future to miniaturize it much more, economics of designing and building defect-free instances are causing a practical halt. It’s already absurdly expensive for a modern process node chip line now. Apple is using 5 nm node now, assuming things scaled down smaller linearly (not a safe assumption) in theory, if we get down to 1 nm process nodes in the future, we can have SoCs with 25 times the density. But, we’re already in quantum mechanical issues territory now.

Interesting comment. What do you see as the current lack of integration between GPUs, PCI-E and Linux?

@@GaryExplains Well... When RPI compute model 4 came. A lot if people tried to get Geforce running on PCI-E on the board. It did see the card and drivers were able to be compiled. However it failed to initialise PCI-E. So there are still work to be done.

I am sure that is true, but that has nothing to do with Armv8 or Armv9. That is about the Pi, the compute module, and an NVIDIA video card etc.

@@GaryExplains I know. But the Arm9, might so advanced, that nvidia will make this happen on PI5 when it is released. It does actually not matter what version of ARM it will work with. As long as it will become possible.

Hmmm... Well the Pi 5 won't be using Arm9 for sure. I would be very surprised if it did. Nvidia isn't interested in getting its hardware to work on the Pi, it has its own Jetson boards which support its GPU just fine.

The new A15 is going to be an iterate update on the A14 so minor efficiency and speed upgrades. The ARM9 will likely be in the iPhone 14 which may also be 3nm and that will probably see a generational leap?

Interest comment Andy. Why do you think that?

@@GaryExplains Rumours suggest an iterate update (TSMC leaks). Reasons: 1/The A15 starts production in May so that may be a tight schedule? 2/Waiting for 3nm generational change. 3/ Waiting for their active cooling solution to be up and running. 4/ COVID times still means that may be slow and steady due to uncertainty in supply and demand?

@@andyH_England 1. The timeline isn't right, Apple has been involved in Armv9 planning from start. 2. 3nm isn't a prerequisite for Armv9. 3. Like above, cooling isn't a prerequisite.

@@GaryExplains Samsung works with ARM and is not expecting ARM9 till next year and th press release suggested later in the year for the first devices, possibly the M2? The M2 may also be 4nm so that would tie in nicely with the ARM9 architecture. The ARM9 architecture claims to offer 30% more performance in two cycles so it is clear that with 4nm the iPhone will need active cooling to allow max usage. We have seen with the SD888 and the iPhone 12 that chips are now too powerful for the smartphone footprint with thermal issues in intensive workloads?

*scalability

@@alvallac2171 I blame fat fingers :-P

eh?

ARM is the architecture, the A15 and M2 will be the system on chip and much like Qualcomm brands them in line with their marketing.

@@andyH_England so the names in cpu like thunder or whatever are sub names for the ones in arm's ?

@@hichamhsr6350 ThunderX3? I think that is the ISA brand name? I suspect when a business designs a new server architecture they brand it? Don’t quote me on that though!

Even if they still use arm v8, they change the architecture of the cpu every year and name changes every year

Yes.

If accurate, this video speaks to that: ukposts.info/have/v-deo/fZWjgmakoYN9qps.html

@@alexcmartens Thanks for that link. So the video is basically accurate until the conclusions at the end. First, it will be Armv9.1, 9.2 etc not A, B, C. The A is for application rather than M (microcontroller) and R (real time). That is why the CPU designs are called Cortex-A. Second, Apple isn't allowed to do its own thing with the instruction set. That is part of the conditions in the license. Apple's cores have to be 100% compliant with the architecture and there is testing involved to ensure that. Thirdly, because of the second point Apple (and others) meet with Arm to help define the direction of the architecture. So the bit about Arm catching is almost right. Apple knows where it is heading and has been using that to help steer Armv9. But it won't and can't do its own thing.

Apple’s hardware keeps it’s price for longer than other PC manufacturers. The old iPhone 5s is still being supported

@@elmersbalm5219 it’s the 6s and SE that are still supported. But yeah they support devices for a very long time

@@chidorirasenganz 5s still receives security updates. 6s receives iOS updates.

@@elmersbalm5219 yeah ios 12 and maybe 11 still get security updates. But by supported I meant new iOS upgrade because those contain new features and such.

@@chidorirasenganz if you want the new features then it’s only fair that you pay for them. Point it that the phone still works an gets security updates. I don’t expect my 10 year old car to get updated features. I do expect spare parts support and safety updates. Germany and the EU are moving to a mandatory 10 year lifetime support for white goods and other technology.

I did that already. See my video on Qualcomm 5nm chips.

Correct and Armv9 is about that instruction set.

@@GaryExplainsah, okay

Lol what

Apple has never used an "X" processor in an iPhone.

The example given of killing 32-bit Macs cannot be compared to ARM8 to ARM9 as there is no record of Apple doing that and I doubt they need to. There were sensible reasons to get rid of legacy 32-bit hardware support which MS refuses to do to their detriment; it allows you to concentrate on a faster and more efficient system. The iPhone 5 (2012) was ARM7 and received iOS 10 (2016) so was supported for five years. I suspect ARM9 iPhones will get six years of OS support and the M1 Macs similar. Of course, they will not stop working at T0, so you could still run your M1 Mac if you keep it for the seventh year, though my experience with Windows laptops suggests it is very rare to last that long. I had a BSOD on my 7th-gen YOGA a few weeks ago!

what's the surprise there? Apple and AMD make their CPUs obsolete every 1-1.5 years. Nothing too surprising....

@@amashaziz2212 That's why i was using a four year old iphone and a 8 year old pc

@@saileshrachapudi8486 similar situation here.

@@amashaziz2212 what do you mean by making CPUs obsolete? Do you want them to release new CPUs once every 5 year?

Unless it was Apple that suggested many of the features of Armv9.

@@GaryExplains Could be - but I believe that Apple wants to distance themselves from ARM Holdings since they want their IP clear from any intellectual entanglements. Apple clearly _could have_ acquired ARM Holdings - certainly there are no national security impediments associated with the Cupertino firm (unlike with nVidea) - but ARM's other clients could've sued Apple claiming that they kept the best stuff for themselves, and tried to horn in on Apple Silicon IP and used various governmental Anti-Trust divisions as a club (which seems to be the popular thing to do nowadays) to force Apple to integrate AS IP into the standard ARM processor technology stream. I think Apple is happy where they are - a parallel technology group observing ARM from a distance with their architectural license - but an independent entity. I don't think Apple minds if ARM borrows ideas from AS since that helps keep the two architectures relatively parallel - but I doubt there's any formal technology exchange. Like I said, avoid intellectual property entanglements or encumbrances. As it is, Apple is in a unique position within the computer/phone/tablet/watch industries: they have exclusive access to what is arguably the most sophisticated processor family in the industry, and other companies producing competing devices cannot simply go down to the Apple Silicon store to buy the same chips so they can duplicate AS price/performance/power efficiency. Microsoft _says_ they'll begin producing their own chips for their Surface Laptops, but saying it and doing it are two entirely different things.

The Apple Silicon IP (the microarchitecture) and the ISA are two very different things from a legal point of view. Apple licenses the ISA from Arm. That is the relationship for Armv8 and nothing changes for Armv9. I know for a fact that Apple has had heavy input into Armv9. You might find this interesting: 9to5mac.com/2021/01/05/magic-of-m1-mac-began-10-years-ago/

@@GaryExplains Well, that's certainly interesting - sorta reminds me of the fact the 64 bit x86 isn't an Intel development but rather AMD's baby. I knew that Apple had the first ARMv8 implementation (for the iPhone 5s), but I was under the impression that Apple had cribbed it from micrographs of an ARM prototype core design. I wasn't aware they were the _instigators and originator_ of AArch64 - almost makes ARM sound like a subcontractor. It also makes sense how wide the Lightning and Firestorm cores are from the A13 and A14 - not happy circumstance but architecturally a design goal. An eight wide processor is certainly something that would be practically impossible to do with a variable length instruction set (like x86), but a tad less impossible with AArch64's fixed instruction length. Sounds like x86 has been in the crosshairs for a while now - well, maybe not in the crosshairs but certainly a candidate for replacement. Every processor maker has been a disappointment to Apple: Motorola 68K, IBM PowerPC, and now Intel x86. All evolutionary paths diverged from what Apple wanted or were incapable of following their product roadmap. Apple obviously wanted to own their processor path forward where there would be no divergence of objective and the silicon could be tailored to be in congruence with future product plans. But eight years - plus a couple of more to get the ball rolling. Kinda scary how long a game Apple has been playing.

It is indeed interesting. You might find my Arm vs Intel video interesting as I cover some of the shared history between Arm and Apple there.

I don't think Armv9 has any connection to raytracing that is a GPU feature, not CPU.

WOW I would never have been able to google that!

They probably will, but that is your fault for buying a 1st Gen product. HOWEVER, be happy with what you have as it's a good laptop for your basic web browsing and light editing.

@@ahlixemus8903 Yes, just buy Intel for iterate upgrades and being years behind the tech curve. With progress comes change and that is the nature of the beast. We would never buy any tech products if we were worried that in a year or so it would be obsolete, apart from Intel, where you know nothing changes much!

@@andyH_England I can disagree or agree with your comment because it all depends on your patience to hold on to something whether it be a phone or laptop. Intel has royally screwed up and I hope that the new CEO can do... SOMETHING. As it stands now, Apple creating their own chips is a wake up call to all chip makers and while it is certainly not from my favorite company, I'm happy they did something unlike Intel. Why the fuck would you buy an Intel chip??? The only reason you should is because you are driven into a corner that you can't go AMD or forbid yourself from buying a Macbook if you are doing more than just web browsing and light video editing.

@@ahlixemus8903 Intel are getting the new 4nm TSMC chips for 2022 so we should see them finally catch up if they can put it altogether in a package. Just how much damage they can avoid in the next year to their bottom line, though I suspect they can afford it even if they do not like losing 33% of their market!

@@andyH_England Intel is certainly to become competitive once more, but the question is if it will be too late or will they become the chip leader once more?

Wait for apple to release their graphics card on a higher spec macbook pro or mac pro. We have seen what they can do on cpu, soon they can do the same thing on gpu as well

*Gary

And how is that connected to Armv9 features like realms?

@@GaryExplains Apple will stay with its own developed security technology and M1 has also already implemented its own virtualization layers which should be close to realms in function. The big difference is - Apple will outsource those functions from the ARM Core into special cores on their SoC, which offers a lot of benefits. The ARM core stays small and efficient which will offer much more headroom when scaling up in count, a special core could be exactly designed for a certain performance per chip and can be completely shut down when you don't need it and because of Unified Memory all special cores can work in parallel without blocking the CPU and last but not least Apple has already invested to much work in their chip design and software developement, so staying and further developing their special cores will be much easier than a restart with ARMv9. Apple will only change to ARMv9 if this architecture offer much more CPU performance, but not because of the new features you mentioned in the video, because all of these problems Apple has already solved with their own technology.

If you noticed Apple was always ahead in terms of the point revision number of the Armv8 architecture. The A14 was Armv8.5. Arm itself was still on Armv8.2. My understanding is that Apple has been one of the key influences of how Armv8 developed and it adopted many of the features as they came out. Armv9 is the same, Apple was very influential in its development and it will adopt it quickly. As for "the ARM core stays small and efficient" then I suggest you go back and have a look at the size of the Firestorm CPU cores, "small and efficient" is not how you would describe them.

@@GaryExplains At second glance I agree with you - Apple will move to ARMv9 - first because Apple can integrate new vector extension and other new functions also in their neural engine / other special cores and and second because Apple needs this performance for their MacPro chips - but my guess is one generation later with A16/M3 with the additional transistor budget of the 3 nm node.

They will attempt to make this useful to crypto miners.

I would assume that any such chip would be called the M2, not the M1X. Apple seems quite consistent in using the "X" name for processors that are enhanced versions of the base model with more cores etc.

@@GaryExplains as much as you say that, Apple has certainly been known to name things with no visible rhyme or reason in a way not done before, if it suited them. I see no reason to guarantee Apple will be certain to maintain any naming convention despite history, if marketing says breaking the past is the right thing. Where was the iPhone 9? Just one example ;)

the area of apple's small core is similar to the cortex A76

@Bilbo Baggins We are talking about software updates BTW, I use Arch.

@Bilbo Baggins My Arch is stable like a stable coin, I update whenever I want, which takes 3 mins in the background, nothing bad happened yet after years of usage, and I am very productive much more than on distros where much of my software just doesn't exist. I am a software developer, I need and want new non snap software, on other platforms like Winshit these software automatically updates too, e.g. via Jetbrains Toolbox, nobody is complaining about that, same software, same devs, same new updates, but on Linux we have unified packaging systems which updates all my software instead of only partially updating and ending up with security issues because you forgot to update your 7z or whatever. Arch does that for me without any problems, which doesn't mean it's perfect, perfect would be categorized package sources with different stability indicators like very stable for desktop environment, bleeding edge for browsers, stable bleeding edge new IDEs, stable kernel, stable whatever should be stable, additional security update channels and so on but that doesn't exist yet. I want my productivity software pretty new, no preview shit versions, but the stable new actual releases, browsers always bleeding edge, distro stuff like DE stable or ultra stable. But my system is stable enough for me like stated above, not a single crash yet except when my hardware completely broke (I am not a lucky person, RAM modules and CPU (5950x) died in January, that swaped out -> 0 crashes, always stable), if it wouldn't be stable, I would not use Arch. :D But you can use whatever you want, you just have to be happy with your choice, or just don't care about it, distro at the end doesn't matter. I am happy with Arch, I don't hop to another while having a completely perfectly working system adapted exactly to my ideas and design, which has all tools I needed; other distros are good too, but I am happy with it. I am using BTRFS with automatic snapshots, in case something bad happens, never had to use one of them yet, but it's all I need. All my software works fine and I have a lot, from video editing, audio editing and so on, 3D modelling, game engines and tons of IDEs, but I mostly use VS Code, and games work fine either, but my GPU currently sucks should not have sold my 2080 Ti. (My English sucks, sry.) If Arch isn't something for you, okay, you decide what fits your needs. For servers I sweared on CentOS, but RedHat... murdered it. RockyLinux is coming. Greetings

Lol 2023 to upgrade Mac? Mac is pointless I hate it. Don’t even bother

I Still have my late 2013 MBP, still runs smoothing but started to show its age a bit.. will upgrade most Likely next year.

Thanks. I have started the clean up!!!

Yes, as are Qualcomm and Samsung and MediaTek. But, this is about CPUs.

*their (possessive) they're = contraction of "THEY aRE" there: for all other uses

@@alvallac2171 Ha, got me!

Quite impossible because M1 is 8w meanwhile phone chips are 5w

@@monkeslayer-km5ho it’s 10W nominal 20W peak. Nowhere is it 8W. The P cluster is rated at 13.8W and E cluster at 1.3W, rest goes to GPU/NPU as needed.

Lol it’s called the a14 , they are the same chips just with a few less cores

@@earnistse4899 also thunderbolt controller

Probably A16-17 although A15 will already be better in Ipc/Single core

nah...not really if they did we would already have a rival to the sd 888.

@@amashaziz2212 They already overtook Qualcomm is chip sales last year. They're now the leading chip manufacturer for Android. In terms of sales yes, but in chip performance they're yet to beat Qualcomm

@@ephraimedits so if u already knew i was talking about chip performance why the HECK bring in chip sales?? How stupid r u huh?

@@amashaziz2212 mediatek doesn't have a chipset in the 888 range It's dimensity 1200, 1000+, 800U all whoop qualcomm's ass

@@DrAdityaReddy wth r u even trying to say dude?