New Arm v9 announced

Krenum

Fully [H]
Joined
Apr 29, 2005
Messages
17,588
tenor.gif
 

defaultluser

[H]F Junkie
Joined
Jan 14, 2006
Messages
13,988
I thought the acquisition got squashed by govt?

No, it's just a bunch of bitching by other big companies that like to whine (mostly Qualcomm), and every government under the sun besides ours doing it's job and investigating the matter.

https://www.forbes.com/sites/bethki...quisition-should-be-approved/?sh=4d1e0a8a45b2

There is no real worry about anyone having a monopoly, as there is nothing stopping Apple from building their own CPUs (all there other ARM licensees already gave-up custom CPU). Additionally, there are like 6 different major GPU vendors in the mobile phone market (besides MALI): Broadcom, AMD, Imagination Tech, Qualcomm, and Apple, so merging two will not do anything negative.

Where the hell are all of you people getting the impression that this merger is already dead, besides pointless paid articles from Qualcomm? Where is there any evidence that they will stop licensing to other companies at current rates?
 
Last edited:

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
This is quite the new feature set, and if ARMv9 can deliver it will definitely have a major inroad to the server, supercomputing, and enterprise markets.
Once NVIDIA completes the merger it is going to give x86-64, and potentially even other ISAs, a run for their money.
 

ChadD

Supreme [H]ardness
Joined
Feb 8, 2016
Messages
5,121
This is quite the new feature set, and if ARMv9 can deliver it will definitely have a major inroad to the server, supercomputing, and enterprise markets.
Once NVIDIA completes the merger it is going to give x86-64, and potentially even other ISAs, a run for their money.
In 20 years people are going to have a hard time remembering how entrenched Intel and x86 seemed.
It is very easy to see why Nvidia is so interested in ARM. I hope our computing platforms stay somewhat open... but yes the next few years are going to be key in how that all unfolds. I really hope we aren't all stuck with disposable kit like Apples M1 stuff. That is the big fear I have though Nvidia I am sure is going to get the go ahead on ARM at some point. I am more worried about Nvidia building sodered on complete ARM/GPU systems and the idea of having upgradable bits at all (not just talking about the CPU) may go out the window. That may sound crazy but I mean if Intel doesn't get their stuff together... in 2-3 years a Nvidia system built to deal with Apple powered by 16 of these V9 cores and even a mid range Nvidia GPU could end up being the true x86 end. Trusting Nvidia to make a upgradable type system where chips from say AMD could be slotted seems unlikely... never mind the potential for Fujitsui Samsung or some other player.

Nvidia could surprise though... Imagine if they actually did spec out a ARM-Socket standard. Perhaps if they owned ARM that possibility increases. I imagine one way for them to argue against gov disallowing their purchase could be... hey we are planning a open platform, we want everyone using ARM. Heck even Qualcomm would have a hard time railing against that.
 

DukenukemX

Supreme [H]ardness
Joined
Jan 30, 2005
Messages
5,402
There is no real worry about anyone having a monopoly, as there is nothing stopping Apple from building their own CPUs (all there other ARM licensees already gave-up custom CPU). Additionally, there are like 6 different major GPU vendors in the mobile phone market (besides MALI): Broadcom, AMD, Imagination Tech, Qualcomm, and Apple, so merging two will not do anything negative.
It certainly will do something negative in the market as Nvidia isn't exactly going to let ARM promote Mali graphics. Mali won't go anywhere with Nvidia controlling ARM. Also, Nvidia isn't exactly friends when it comes to relations with others. Did we forget about Nvidia's Geforce Partner Program that Kyle called them out on? Not Nvidia's first time being problematic with other companies. Remember Linus Torvalds fuck you? Microsoft Corp declining to comment on its relationship with Nvidia after the shit they went through with Nvidia? Apple still hates Nvidia from the whole GeForce 9600M fiasco and the lack of support Nvidia gave. That's why you haven't seen any Nvidia products in Apple laptops.

That's just with Nvidia and other businesses, but there's issues with consumers as well. Anyone remember Nvidia's GameWorks and how it made games worse but less worse for Nvidia cards? Assassin's Creed loses DX10.1 support which benefited ATI a lot. How about the whole 3.5GB of 4GB GTX 970 fiasco? More recently they got pissed off about the lack of Ray-Tracing benchmarks from Hardware Unboxed and was going to take away giving them GPU's. I can go on and on about how Nvidia handles things. Do you really want a company like Nvidia, controlling something as important as ARM? Nvidia might be what ARM needs to move forward but it will be through propriety hell and legal threats with closed standards. I can already see ARM with stuff like PhysX and Cuda crap that nobody else can use without a licensing fee. I can see companies dumping Nvidia's ARM for MIPs or RISC-V.


 

defaultluser

[H]F Junkie
Joined
Jan 14, 2006
Messages
13,988
It certainly will do something negative in the market as Nvidia isn't exactly going to let ARM promote Mali graphics. Mali won't go anywhere with Nvidia controlling ARM. Also, Nvidia isn't exactly friends when it comes to relations with others. Did we forget about Nvidia's Geforce Partner Program that Kyle called them out on? Not Nvidia's first time being problematic with other companies. Remember Linus Torvalds fuck you? Microsoft Corp declining to comment on its relationship with Nvidia after the shit they went through with Nvidia? Apple still hates Nvidia from the whole GeForce 9600M fiasco and the lack of support Nvidia gave. That's why you haven't seen any Nvidia products in Apple laptops.

That's just with Nvidia and other businesses, but there's issues with consumers as well. Anyone remember Nvidia's GameWorks and how it made games worse but less worse for Nvidia cards? Assassin's Creed loses DX10.1 support which benefited ATI a lot. How about the whole 3.5GB of 4GB GTX 970 fiasco? More recently they got pissed off about the lack of Ray-Tracing benchmarks from Hardware Unboxed and was going to take away giving them GPU's. I can go on and on about how Nvidia handles things. Do you really want a company like Nvidia, controlling something as important as ARM? Nvidia might be what ARM needs to move forward but it will be through propriety hell and legal threats with closed standards. I can already see ARM with stuff like PhysX and Cuda crap that nobody else can use without a licensing fee. I can see companies dumping Nvidia's ARM for MIPs or RISC-V.


And?

By comparison, the shit NVIDIA has done, Apple's lock-in and 32-bit deletion makes them look like saints. And Qualcomm's forced OEM's cross-licensing contracts by force of Lawsuit (plus holding up the entire Android OS with their shit driver update schedule plus killing TWO internal CPU teams)P

NVIDIA sucks in their own little island, ut of the other two companies I just mentioned, their shit stinks a lot less. When you're talking about a company that can actually afford arm, the list of nice companies is nowhere to be found!
 
Last edited:

DukenukemX

Supreme [H]ardness
Joined
Jan 30, 2005
Messages
5,402
And?

By comparison, the shit NVIDIA has done, Apple's lock-in and 32-bit deletion makes them look like saints. And Qualcomm's forced OEM's cross-licensing contracts by force of Lawsuit (plus holding up the entire Android OS with their shit driver update schedule plus killing TWO internal CPU teams)P

NVIDIA sucks in their own little island, ut of the other two companies I just mentioned, their shit stinks a lot less. When you're talking about a company that can actually afford arm, the list of nice companies is nowhere to be found!
Plenty of shit companies to go around but I'm pretty sure Nvidia is nearly king of the kill on this one. Especially when you need ARM to be as open as possible.
 

defaultluser

[H]F Junkie
Joined
Jan 14, 2006
Messages
13,988
Plenty of shit companies to go around but I'm pretty sure Nvidia is nearly king of the kill on this one. Especially when you need ARM to be as open as possible.
And although you mention Mali as an obvious target, do you know how much less they can charge fir it versus CPUs? The Royalty rates are roughly half (otherwise, you will lose the design win to Imagination Tech.)

een%20Shot%202013-06-27%20at%2011.40.03%20PM_575px.png

Now, when HALF OF ALL PHONE CHIPS (Qualcomm + Apple) already have their own proprietary GPU, ARM is flailing to make money off more -and-more complex GPUs, it would be easier for NVIDIA to merge the two divisions (rebrand Mali as entry-level GTX). Acquire the best talent, then make the redundancies disappear (and put the increased profits towards more advanced ARM CPU designs).
\
bOTH COMPANIES GAIN FROM THE MERGER (EVEN IF SOME DESIGN JOBS WILL BE LOST), and on the unlikely chance they have an advantage, it would appear in the next rtx.
 
Last edited:

ChadD

Supreme [H]ardness
Joined
Feb 8, 2016
Messages
5,121
x86 is done for this time!

Linux is going to pull ahead on the desktop this time!

Duke Nukem Forever will be out this year!

Stop me if you have heard this before.

Well its easy to not see what is happening to x86 if your a desktop / gamer user. ARM hasn't done anything to go after that market at this point. If I told you 3 years ago that half of Amazons cloud sales would be for ARM powered instances by 2021.... would you have believed me ?
Amazon just started spinning up graviton 2 instances and are claiming 55% better price to performance vs their x86 instances. They got to a 50/50 split with 35% better price to performance out of graviton 1. I'll tell you now that in another 2 years Amaozn will be at a min of 80/20 and I wouldn't be shocked if it more like x86 instances dropping to single digits running a handful of very old code.

ARM has now taken the super computer crown... and when that happened people said oh this will be short lived cause the US has this Intel powered machine coming up. Well seems like Intel is having a hard time keeping that on schedule but the US has dropped their lawsuit so hopefully its back on track (or Intel just dropped their pants on the pricing scheme).

I won't make the usual ARM powers more devices argument. I believe its valid. However realistically the reality of covid lockdowns have breathed some new life into the desktop platform. With Nvidia looking to seal their ARM deal... and now that we see the next gen ARM cores and where they are heading in terms of synergies. Yes it seems pretty obvious that Apples M1 isn't going to be an oddity. Apple hits the desktop themselves later this year that is known. I believe Nvidia is applying all the pressure they can to get the ARM thing locked down as quickly as possible. The first real push for ARM on the desktop is coming very soon, and making it the last major market they go after was probably wise. As I said before my fear is they are going to try and kill off the entire open platform idea.... its easy to say it will never happen. However if you give most consumers the choice of a x86+GPU or a locked down desktop that is 3/4 the cost for 30% more performance? The masses may not care all that much if the CPU RAM and GPU are all soldered to the same board.
 

Aurelius

2[H]4U
Joined
Mar 22, 2003
Messages
3,162
I won't make the usual ARM powers more devices argument. I believe its valid. However realistically the reality of covid lockdowns have breathed some new life into the desktop platform. With Nvidia looking to seal their ARM deal... and now that we see the next gen ARM cores and where they are heading in terms of synergies. Yes it seems pretty obvious that Apples M1 isn't going to be an oddity. Apple hits the desktop themselves later this year that is known. I believe Nvidia is applying all the pressure they can to get the ARM thing locked down as quickly as possible. The first real push for ARM on the desktop is coming very soon, and making it the last major market they go after was probably wise. As I said before my fear is they are going to try and kill off the entire open platform idea.... its easy to say it will never happen. However if you give most consumers the choice of a x86+GPU or a locked down desktop that is 3/4 the cost for 30% more performance? The masses may not care all that much if the CPU RAM and GPU are all soldered to the same board.
Threads like this highlight one of the perpetual problems with tech enthusiast forums: this tendency to assume gaming PCs are the center of the universe, and that no computing platform can succeed without gamers' blessing. But devoted gamers only represent a fraction of the market, even if you exclude office PCs. And we might be on the cusp of seeing just how limited gamers' reach really is. Apple won't have a huge effect unless its next Macs really take off, but I wouldn't rule out a scenario in the next several years where x86 home PCs end up as outliers, albeit significant ones with a huge legacy installed base.
 

DukenukemX

Supreme [H]ardness
Joined
Jan 30, 2005
Messages
5,402
And although you mention Mali as an obvious target, do you know how much less they can charge fir it versus CPUs? The Royalty rates are roughly half (otherwise, you will lose the design win to Imagination Tech.)

View attachment 343835

Now, when HALF OF ALL PHONE CHIPS (Qualcomm + Apple) already have their own proprietary GPU, ARM is flailing to make money off more -and-more complex GPUs, it would be easier for NVIDIA to merge the two divisions (rebrand Mali as entry-level GTX). Acquire the best talent, then make the redundancies disappear (and put the increased profits towards more advanced ARM CPU designs).
\
bOTH COMPANIES GAIN FROM THE MERGER (EVEN IF SOME DESIGN JOBS WILL BE LOST), and on the unlikely chance they have an advantage, it would appear in the next rtx.
There was a rumor that ARM's mali was going to enter the PC market, which means it would have competed with Nvidia. The reason ARM went for sale was because they invested too much into Internet of Things (IoT). Turns out people didn't need toasters and refrigerators that were connected to the internet. If ARM went after the PC graphics market with Mali then things would be different.

If I told you 3 years ago that half of Amazons cloud sales would be for ARM powered instances by 2021.... would you have believed me ?
I'm surprised it wasn't done sooner. Amazon using ARM for a cloud server isn't anything important. Apple's M1 is far more important for ARM's future.
Amazon just started spinning up graviton 2 instances and are claiming 55% better price to performance vs their x86 instances. They got to a 50/50 split with 35% better price to performance out of graviton 1. I'll tell you now that in another 2 years Amaozn will be at a min of 80/20 and I wouldn't be shocked if it more like x86 instances dropping to single digits running a handful of very old code.
Depends on which x86 you compare it with. With Intel I can believe that since Intel is still stuck on 14nm. AMD on the other hand... People don't understand how badly Intel fucked up being stuck with 14nm. Amazon's AWS is also on 7nm.
However if you give most consumers the choice of a x86+GPU or a locked down desktop that is 3/4 the cost for 30% more performance? The masses may not care all that much if the CPU RAM and GPU are all soldered to the same board.
That's basically console gaming vs PC, and you can still see that people prefer their open platform much more.

Threads like this highlight one of the perpetual problems with tech enthusiast forums: this tendency to assume gaming PCs are the center of the universe, and that no computing platform can succeed without gamers' blessing. But devoted gamers only represent a fraction of the market, even if you exclude office PCs. And we might be on the cusp of seeing just how limited gamers' reach really is. Apple won't have a huge effect unless its next Macs really take off, but I wouldn't rule out a scenario in the next several years where x86 home PCs end up as outliers, albeit significant ones with a huge legacy installed base.
Posts like this is the reason why Linux still hasn't taken over the desktop market. Gamers are why PC's mainly get faster. If gamers weren't involved in PC then a tablet is all you'd need to do your daily tasks. Obviously not everyone is using a tablet for everything, so there's a reason for this. There's a reason why a GPU went from playing Quake 3 to rendering an OS desktop to rendering video encoding. We are the reason why anyone spends money to make faster hardware. We are the reason why Doom outsold Windows 95. We are also the reason why nobody takes Apple serious.

Think of it like what porn did for DVD. If you want to get anywhere in this market then you need gamers. ARM's been out for decades and can do basically anything x86 can do now, but it doesn't have the software. It doesn't have an open platform. More importantly, it doesn't have any games... besides mobile. Fuck mobile games. If ARM is to get anywhere on the desktop market then it needs to open up and standardize, just like our modern day IBM compatibles.

 
Last edited:

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
Posts like this is the reason why Linux still hasn't taken over the desktop market. Gamers are why PC's mainly get faster. If gamers weren't involved in PC then a tablet is all you'd need to do your daily tasks. Obviously not everyone is using a tablet for everything, so there's a reason for this. There's a reason why a GPU went from playing Quake 3 to rendering an OS desktop to rendering video encoding. We are the reason why anyone spends money to make faster hardware. We are the reason why Doom outsold Windows 95. We are also the reason why nobody takes Apple serious.

Think of it like what porn did for DVD. If you want to get anywhere in this market then you need gamers. ARM's been out for decades and can do basically anything x86 can do now, but it doesn't have the software. It doesn't have an open platform. More importantly, it doesn't have any games... besides mobile. Fuck mobile games. If ARM is to get anywhere on the desktop market then it needs to open up and standardize, just like our modern day IBM compatibles.
All of this - exactly this, period.
Once a standard is met, a true contender to x86-64 will have been fully realized.

You are right, Apple started this with M1, but is still too proprietary and/or vendor-locked to run anything outside of MacOS, yet.
Once a mainstream and/or open bootup and platform standard for ARM are both met, it will be truly game over for Intel's x86-64, not counting AMD unless they somehow stop innovating the way they have for the last 4 years.

Their latest CPUs are a joke, even when directly compared to their own last generation CPUs, and it seems like the new CEO, and management shakeup, has done little to improve their situation.
Intel needs to get its proverbial shit together, and soon.
 
Last edited:

ChadD

Supreme [H]ardness
Joined
Feb 8, 2016
Messages
5,121
There was a rumor that ARM's mali was going to enter the PC market, which means it would have competed with Nvidia. The reason ARM went for sale was because they invested too much into Internet of Things (IoT). Turns out people didn't need toasters and refrigerators that were connected to the internet. If ARM went after the PC graphics market with Mali then things would be different.


I'm surprised it wasn't done sooner. Amazon using ARM for a cloud server isn't anything important. Apple's M1 is far more important for ARM's future.

Depends on which x86 you compare it with. With Intel I can believe that since Intel is still stuck on 14nm. AMD on the other hand... People don't understand how badly Intel fucked up being stuck with 14nm. Amazon's AWS is also on 7nm.

That's basically console gaming vs PC, and you can still see that people prefer their open platform much more.


Posts like this is the reason why Linux still hasn't taken over the desktop market. Gamers are why PC's mainly get faster. If gamers weren't involved in PC then a tablet is all you'd need to do your daily tasks. Obviously not everyone is using a tablet for everything, so there's a reason for this. There's a reason why a GPU went from playing Quake 3 to rendering an OS desktop to rendering video encoding. We are the reason why anyone spends money to make faster hardware. We are the reason why Doom outsold Windows 95. We are also the reason why nobody takes Apple serious.

Think of it like what porn did for DVD. If you want to get anywhere in this market then you need gamers. ARM's been out for decades and can do basically anything x86 can do now, but it doesn't have the software. It doesn't have an open platform. More importantly, it doesn't have any games... besides mobile. Fuck mobile games. If ARM is to get anywhere on the desktop market then it needs to open up and standardize, just like our modern day IBM compatibles.

I would agree with much of what you said. But where I disagree is where you say this is like Console vs PC. No its really not. I didn't say 30% LESS performance. I said 30% MORE. Make no mistake when Nvidia comes swinging for the desktop market they are going to be doing it with a 12-16 core performance ARM core that is going to spank x86... and I would suspect probably couple it with a standard PCIX slot for a GPU. (the question of that slot supporting anything but Nvidia GPUs is in question). I have no doubt when they come it won't be a take this locked down Nvidia game console.... it will be this is a Nvidia gaming station, and even gamers will be swung when its simply spanks anything Intel and AMD are offering.

I hope Nvidia goes full open and standardizes a ARM socket... but its looking pretty unlikely.
 

UnknownSouljer

Supreme [H]ardness
Joined
Sep 24, 2001
Messages
7,028
That's basically console gaming vs PC, and you can still see that people prefer their open platform much more.
This is the only part of your post I have anything in particular to speak against. I don’t think you know even where most consumer PC sales are going then. Which is of course the laptop market. Things changed a bit because of COVID19, but honestly more people still prefer a laptop over a desktop PC. This is a space that there are zero disadvantages to ARM (as in, “being open” has no real advantages - people buy laptops like appliances). And as ChadD noted, if it’s faster and costs less it becomes a no brainer scenario.

Windows is already losing market share due to Chromebooks (it’s now the number 2 OS, passing macOS despite macOS also having a good growth year). I see it’s a very direct future in which Google replaces their Intel processors for ARM, moves over their software stack, and can advertise a product that is faster, less expensive, with better battery life than competing windows machines. And people primarily doing office/browsing tasks will happily move over - as they already are.
 

1_rick

[H]ard|Gawd
Joined
Feb 7, 2017
Messages
1,704
they are going to be doing it with a 12-16 core performance ARM core that is going to spank x86
Right up until it runs into an app that wasn't ported to ARM, and then performance falls off a cliff, just like with the Windows ARM tablets.
 
  • Like
Reactions: travm
like this

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
Right up until it runs into an app that wasn't ported to ARM, and then performance falls off a cliff, just like with the Windows ARM tablets.
Except, if done right, that isn't really the case, especially with MacOS 11 and Rosetta 2, which is a translation layer rather than emulation, thus performance doesn't suffer nearly as much.
Modern apps will need to be ported or updated to support multiple ISAs (or be ISA-agnostic), depend on emulation/translation, or if proprietary enough then needing to hold onto legacy equipment to perform the task for however long it can.
 

DukenukemX

Supreme [H]ardness
Joined
Jan 30, 2005
Messages
5,402
You are right, Apple started this with M1, but is still too proprietary and/or vendor-locked to run anything outside of MacOS 11, yet.
Once a mainstream and/or open bootup and platform standard for ARM are both met, it will be truly game over for Intel's x86-64,
Not really. One major step would be to make an open platform for ARM just like x86's modern day IBM compatible PCs. You still need the software, which ARM does have a lot of software but nothing on the scale of x86 Windows. Did you know there's an ARM Windows? Nobody knows about it because nobody in their right mind would use an ARM based Windows laptop over an x86, because of the sheer amount of software available on it. Microsoft can't take down x86 Windows. Linux can't get people to switch and make x86 Linux apps. Steam had to make Proton so that x86 Windows games would run on x86 Linux, because nobody was porting their games over to x86 Linux.

When you make a new CPU platform, you do have a chicken or egg situation. Which came first, the software developers or the users? Trick question, you need both at the same thing. This is why Apple made sure they have decent x86 emulation because nobody is going to develop apps for ARM MacOSX when nobody owns them. Consequently no user is going to go ARM MacOSX when there's no apps to use on it. It'll take Apple years just to have the same level of software like they have now on x86. Maybe even 10 years.
not counting AMD unless they somehow stop innovating the way they have for the last 4 years.
AMD is already competitive against ARM. They may not be using 5nm or 3nm like some SoC's you find on cell phones and tablets but AMD has proven that x86 still has a lot of life left. Also I'm kinda excited about AVX-512 on AMD Zen. I love hearing what emulator authors can do with AVX-512.
Their latest CPUs are a joke, even when directly compared to their own last generation CPUs, and it seems like the new CEO, and management shakeup, has done little to improve their situation.
Intel needs to get its proverbial shit together, and soon.
Intel fucked up because they were dependent on their superior manufacturing. That and Sandy Bridge was a huge boost of performance compared to anything else before it. Then Intel got lazy and didn't want to invest in their manufacturing, which is why everyone else is using 7nm or less while Intel is still stuck on 2014 technology. As great as Sandy Bridge was, that was introduced in 2011. Since Sandy Bridge, Intel hasn't done a major redesign for nearly 10 years. AMD introduced the Athlon 64 in 2003 and were using that architecture until Bulldozer which was 2011. That didn't work out well for AMD, so imagine how badly this is working for Intel. You don't want to use an architecture for too long, otherwise you risk having your competitor out perform you.

This isn't indicative of x86's performance, just Intels. When AMD released the Athlon 64, it was destroying Intel in benchmarks because Netburst=LoL. Yet everyone who compared Apple's PowerPC to x86, they would point out Intel. This is the same situation.
 

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
Not really. One major step would be to make an open platform for ARM just like x86's modern day IBM compatible PCs. You still need the software, which ARM does have a lot of software but nothing on the scale of x86 Windows. Did you know there's an ARM Windows? Nobody knows about it because nobody in their right mind would use an ARM based Windows laptop over an x86, because of the sheer amount of software available on it. Microsoft can't take down x86 Windows. Linux can't get people to switch and make x86 Linux apps. Steam had to make Proton so that x86 Windows games would run on x86 Linux, because nobody was porting their games over to x86 Linux.
Exactly, you hit the nail on the head with the primary issue and what is needed.
Oh yes, I do remember Windows RT from the mid-2010s with it's lame x86-to-ARM emulation and how Microsoft dropped support far too soon for it due to it being such a financial detriment.
When you make a new CPU platform, you do have a chicken or egg situation. Which came first, the software developers or the users? Trick question, you need both at the same thing. This is why Apple made sure they have decent x86 emulation because nobody is going to develop apps for ARM MacOSX when nobody owns them. Consequently no user is going to go ARM MacOSX when there's no apps to use on it. It'll take Apple years just to have the same level of software like they have now on x86. Maybe even 10 years.
That's true, and Apple has been working at getting a software-base going for ARM since the original iPhone back in 2007, so by 2020 they had over a decade worth of software, and an OS ecosystem they had tight control over with little between the mobile and desktop variants, at least on the back-end.
Microsoft did not have this with Windows RT, was relying on emulation, the hopes that developers would catch on (they didn't), and had the disadvantage of using the then 32-bit ARM CPU/SoC solutions on the market.
AMD is already competitive against ARM. They may not be using 5nm or 3nm like some SoC's you find on cell phones and tablets but AMD has proven that x86 still has a lot of life left. Also I'm kinda excited about AVX-512 on AMD Zen. I love hearing what emulator authors can do with AVX-512.
Agreed, AMD has really shown that x86/x86-64 can continue to perform, well beyond what I thought was even capable of with it.
Today, I wouldn't touch an x86-64 CPU that wasn't AMD.
Intel fucked up because they were dependent on their superior manufacturing. That and Sandy Bridge was a huge boost of performance compared to anything else before it. Then Intel got lazy and didn't want to invest in their manufacturing, which is why everyone else is using 7nm or less while Intel is still stuck on 2014 technology. As great as Sandy Bridge was, that was introduced in 2011. Since Sandy Bridge, Intel hasn't done a major redesign for nearly 10 years.
I do remember that as well, and how the HD2000 iGPU was quite the game-changer of 2011, and thinking how good those CPUs and iGPUs really were compared nearly anything else on the market.
Intel only had a ~25% IPC uplift from Sandy Bridge in 2011 to Kaby Lake in 2016, which isn't great, and considering how they forced nearly all end-users into quad-core CPUs for nearly a decade, it did not give anyone confidence in their products.
AMD introduced the Athlon 64 in 2003 and were using that architecture until Bulldozer which was 2011. That didn't work out well for AMD, so imagine how badly this is working for Intel. You don't want to use an architecture for too long, otherwise you risk having your competitor out perform you.

This isn't indicative of x86's performance, just Intels. When AMD released the Athlon 64, it was destroying Intel in benchmarks because Netburst=LoL. Yet everyone who compared Apple's PowerPC to x86, they would point out Intel. This is the same situation.
haha, I remember both of those releases very well, too!
AMD was definitely the performance leader of the mid-2000s until Intel released Conroe in early 2007.

At the time, clock-for-clock, IBM's PowerPC offerings were more powerful overall compared to then x86 CPUs, but tended to be clocked slower due to high TDP and thermals, thus giving similar performance.
By 2006, though, both AMD and Intel had CPUs that could outperform what IBM (and Motorola) could deliver, and the writing was on the wall for PowerPC with Apple; cost and lack of revenue and a lower customer-base were also big factors for IBM to drop out of the market with Apple, and vice versa.

Yeah, Netburst, I remember "dual-core" Pentium D CPUs with a near 250 watt TDP.
Intel's latest offerings are starting to feel that way all over again, only this time with nearly a 300 watt TDP.

If only AMD had Intel's resources and assets, then x86/x86-64 would much more cemented for the following decade.
I can't imagine why NVIDIA wants to invest into ARM so badly. :D

I had always hoped that AMD's ARM CPUs would have taken off further than they did circa 2013.
Perhaps we will see a future AMD ARM CPU or SoC in the future with mainstream usage.
 

1_rick

[H]ard|Gawd
Joined
Feb 7, 2017
Messages
1,704
Except, if done right, that isn't really the case, especially with MacOS 11 and Rosetta 2, which is a translation layer rather than emulation, thus performance doesn't suffer nearly as much.
Modern apps will need to be ported or updated to support multiple ISAs (or be ISA-agnostic), depend on emulation/translation, or if proprietary enough then needing to hold onto legacy equipment to perform the task for however long it can.
I'm thinking also of all those businesses that are still running mainframe apps. Now sure, they can run on terminals in an ARM machine just as easily as anything else but I bet the original ERP or whatever will never be natively ARM, or they would've already ported it to something else.
 

Lakados

2[H]4U
Joined
Feb 3, 2014
Messages
3,796
There was a rumor that ARM's mali was going to enter the PC market, which means it would have competed with Nvidia. The reason ARM went for sale was because they invested too much into Internet of Things (IoT). Turns out people didn't need toasters and refrigerators that were connected to the internet. If ARM went after the PC graphics market with Mali then things would be different.


I'm surprised it wasn't done sooner. Amazon using ARM for a cloud server isn't anything important. Apple's M1 is far more important for ARM's future.

Depends on which x86 you compare it with. With Intel I can believe that since Intel is still stuck on 14nm. AMD on the other hand... People don't understand how badly Intel fucked up being stuck with 14nm. Amazon's AWS is also on 7nm.

That's basically console gaming vs PC, and you can still see that people prefer their open platform much more.


Posts like this is the reason why Linux still hasn't taken over the desktop market. Gamers are why PC's mainly get faster. If gamers weren't involved in PC then a tablet is all you'd need to do your daily tasks. Obviously not everyone is using a tablet for everything, so there's a reason for this. There's a reason why a GPU went from playing Quake 3 to rendering an OS desktop to rendering video encoding. We are the reason why anyone spends money to make faster hardware. We are the reason why Doom outsold Windows 95. We are also the reason why nobody takes Apple serious.

Think of it like what porn did for DVD. If you want to get anywhere in this market then you need gamers. ARM's been out for decades and can do basically anything x86 can do now, but it doesn't have the software. It doesn't have an open platform. More importantly, it doesn't have any games... besides mobile. Fuck mobile games. If ARM is to get anywhere on the desktop market then it needs to open up and standardize, just like our modern day IBM compatibles.

I know a lot of classrooms developing games to natively run on the Pi4's, nothing fancy things that basically look like SNES clones, but the kids are learning on them, and those of them that continue with it will go on to make bigger more awesome things but like all the big companies know "hook them while they're young". It's why Microsoft and Adobe basically give their software to the Education system, but if they are all learning on Arm, and they go onto university and continue with ARM they will go out into the workforce having worked on ARM for close to a decade at that point, it may take 10-15 years but if the teaching trend continues and that hardware stays as cheap as it is and as well supported and documented as it is with the community continuing to grow it is an eventuality that nobody can ignore.
 

Aurelius

2[H]4U
Joined
Mar 22, 2003
Messages
3,162
Posts like this is the reason why Linux still hasn't taken over the desktop market. Gamers are why PC's mainly get faster. If gamers weren't involved in PC then a tablet is all you'd need to do your daily tasks. Obviously not everyone is using a tablet for everything, so there's a reason for this. There's a reason why a GPU went from playing Quake 3 to rendering an OS desktop to rendering video encoding. We are the reason why anyone spends money to make faster hardware. We are the reason why Doom outsold Windows 95. We are also the reason why nobody takes Apple serious.

Think of it like what porn did for DVD. If you want to get anywhere in this market then you need gamers. ARM's been out for decades and can do basically anything x86 can do now, but it doesn't have the software. It doesn't have an open platform. More importantly, it doesn't have any games... besides mobile. Fuck mobile games. If ARM is to get anywhere on the desktop market then it needs to open up and standardize, just like our modern day IBM compatibles.

This... proves my point. You think the gaming PC is the center of the universe, and everyone owes everything to it.

Gaming PCs drive some things: GPUs, certain kinds of CPUs, high refresh rate displays, that sort of thing. But those only represent a slice of the PC industry. Laptops lead PC sales, and most of those are built around efficiency, not raw power. Work-oriented PCs are a large chunk of the market; it's safe to say most PC users aren't buying high-end or even mid-range GPUs. There's a bit of a trickle-down effect, but gamers aren't going to have much influence on ultraportable laptops with 15W chips.

People take Apple plenty seriously (just look at Intel's panicky ad campaign) — it's not dominant for a number of other reasons, including Microsoft's success in carving out a monopoly years ago (and the resulting wider catalog of Windows app), missteps in the 1990s and an emphasis on premium computers that limits the audience. It's difficult to sell a $1,000 MacBook Air to a middle-class Chinese person for whom $400 is a huge outlay.

ARM does face challenges, but the very argument we're making is that it's at an inflection point where it's more than a novelty, where developer support is picking up. I don't think the issue is so much opening up and standardization as just having more than one good ARM-based platform... ARM on Windows could fare so much better if Microsoft and Qualcomm could figure out how to support it!

(Also, mobile games are just fine... remember, Among Us and Fortnite arguably owe large chunks of their success to mobile.)
 

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
I'm thinking also of all those businesses that are still running mainframe apps. Now sure, they can run on terminals in an ARM machine just as easily as anything else but I bet the original ERP or whatever will never be natively ARM, or they would've already ported it to something else.
Mainframes (assuming modern units) are specialized hardware with massive amounts of I/O that no other ISA or platform can touch.
Not saying the applications couldn't be ported or migrated, but if the performance and/or I/O of the mainframe is needed for the task, no other hardware out there is going to be able to handle that type of job or load.

There is a reason IBM has had a stranglehold on the mainframe industry for so many decades, and I doubt ARM is going to even phase it.
That's like saying this new revolutionary car is going to impact fighter jets - like, they are both technically vehicles, but the car won't ever be able to do what the fighter jet can, no matter how efficient it gets... unless we get flying cars, then maybe.
 

cybereality

Supreme [H]ardness
Joined
Mar 22, 2008
Messages
6,959
ARM CPUs for desktop would be cool, if they still let you build your own. Also open access to build/distribute apps (exes or whatever they call them). None of that UWP locked to Windows Store BS.

Also, it could be a problem for legacy apps if the emulator isn't good enough.

Kind of funny, I used to work IT and back around 2001 I worked at this place and their payroll system was on DOS. I offered to write them a new app, but they said it worked fine. They had the DOS server on an old computer in the back of the office.

Sure there are a lot of legacy proprietary apps that may not work with the emulator, or companies would be hesitant to risk it. Apple likes making bold moves, but I can't see Microsoft abandoning their legacy unless the emulator was 100%.
 

ChadD

Supreme [H]ardness
Joined
Feb 8, 2016
Messages
5,121
The other elephant that hasn't been brought up is Microsoft hardware. Yes we all agree MS windows ARM is currently crap and Qualcomm is at least right now not the Desktop ARM champion they would love to be.

The rumor mill is pretty strong that MS is designing their own ARM chip. I still believe Nvidia fully intends to be THE desktop ARM supplier and that leather jacket man has always wanted to be king of the CPUs. Intel screwed him out of his plans years ago and chased NV out of the motherboard chipset business. Buying arm and becoming to ARM what Intel is to x86... and then rolf stomping Intel is the wet dream Jensen has been having for a decade.

Still it looks like MS is ready to take ARM seriously now that it looks like Apple is getting ready to bring some serious pain. Like it or not M1 is selling and selling very well. If the rumors of Apple having a big core bump for a desktop part are even half true... the race to compete is going to lead to some very crazy things over the next few years. I fully expect by the end of the year MS will be talking about a MS CPU.... and Nvidia will be talking about NVs own (new) CPU system as well.

I don't think Nvidia is smart enough to go all in... but if it they wanted to shut down all opposition to their ARM deal. All they have to do is build out a ARM socket spec. I appreciated x86s open arch.... but lets all be honest the competition level blows. Intel or AMD ? Its worse then American politics (sorry). If Nvidia was smart enough to build a ARM-Socket1... and we could buy ARM mobos, and slot CPUs from Nvidia/AMD/Qualcomm/Samsung/Fujitsu/Ampere/Amazon we would have real competition in the CPU market. Same instruction set tons of competition on implementation and design. Now like I say I doubt Nvida is either smart enough... OR confident enough to do that. More likely they are going to build their own locked down things and not try to make it easy for some third party outsider like Fujitsu to build some socket home version of the A64FX, or worse Amazon to start retailing Graviton chips. But I mean as an enthusiast you have to love the idea.... it would be a golden age of CPUs. It be better then the glory days of the 90s with the outlaw x86 builders. :)
 

ManofGod

[H]F Junkie
Joined
Oct 4, 2007
Messages
12,210
No, it's just a bunch of bitching by other big companies that like to whine (mostly Qualcomm), and every government under the sun besides ours doing it's job and investigating the matter.

https://www.forbes.com/sites/bethki...quisition-should-be-approved/?sh=4d1e0a8a45b2

There is no real worry about anyone having a monopoly, as there is nothing stopping Apple from building their own CPUs (all there other ARM licensees already gave-up custom CPU). Additionally, there are like 6 different major GPU vendors in the mobile phone market (besides MALI): Broadcom, AMD, Imagination Tech, Qualcomm, and Apple, so merging two will not do anything negative.

Where the hell are all of you people getting the impression that this merger is already dead, besides pointless paid articles from Qualcomm? Where is there any evidence that they will stop licensing to other companies at current rates?

Nope, it's only Nvidia, no danger of a monopoly here. /s All we have to do is ignore their history and believe there is nothing to see here.
 

travm

Gawd
Joined
Feb 26, 2016
Messages
863
Except, if done right, that isn't really the case, especially with MacOS 11 and Rosetta 2, which is a translation layer rather than emulation, thus performance doesn't suffer nearly as much.
Modern apps will need to be ported or updated to support multiple ISAs (or be ISA-agnostic), depend on emulation/translation, or if proprietary enough then needing to hold onto legacy equipment to perform the task for however long it can.
since when have large corps ever made a point of doing something "right".... That might be the primary driver for X86 being what it is, its got tools for everything.
 

DejaWiz

Fully [H]
Joined
Apr 15, 2005
Messages
20,989
Through all of this, I think Apple has recently dropped the ball when they had a golden opportunity to propel ARM. New M1 seems quite stout, as well as significantly cheaper to make and implement than x86. Apple probably should have used those facts to their advantage with a massive advertising smear campaign against PC to educate the masses and lower their base MSRPs, but they didn't. They are still too proud of their pricing, and that's why Apple has only about 8% of the personal computer global hardware market and around 10% of the OS global market share. Windows is around 80%, globally.
I consider Apple's computer hardware and OS market share numbers as niche. A strong niche, but still niche.

Technology is cheap to manufacture en masse nowadays...Apple is putting cheap hardware in cheap to make packages (just like every other hardware company) and slapping on some of the most massive price tags while marketing them as trendy home decor and fashion accessories. That's a facepalm that Apple has been eliciting for well over a decade.
 

1_rick

[H]ard|Gawd
Joined
Feb 7, 2017
Messages
1,704
Still it looks like MS is ready to take ARM seriously now that it looks like Apple is getting ready to bring some serious pain. Like it or not M1 is selling and selling very well.
I'm just trying to picture all of those middle-aged women in the office, the ones who can barely use a computer and freak out any time anything changes at all--they're the ones that have notepads full of exactly every mouse click and keypress to do their jobs, and heaven help if they miss a step one time--switching from Windows to Apple and wondering how well that'd go.
 

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
Technology is cheap to manufacture en masse nowadays...Apple is putting cheap hardware in cheap to make packages (just like every other hardware company) and slapping on some of the most massive price tags while marketing them as trendy home decor and fashion accessories. That's a facepalm that Apple has been eliciting for well over a decade.
Apple didn't become a $1 trillion+ megacorp because of their low low prices.
Marketing and brand recognition, along with high pricing, has worked for them as a "premium" product provider for over 20 years now, so why now would they change the game?

Doing so would only make their market more competitive, which is the opposite of what they want to happen, and even purely from a business standpoint, it would be a foolish move.
Not sure how that is a facepalm, as obviously it is working quite well for them - that facepalm should go to Microsoft for being late to the game, stifling innovation, and dropping the ball for nearly 20 years.
 

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,003
I'm just trying to picture all of those middle-aged women in the office, the ones who can barely use a computer and freak out any time anything changes at all--they're the ones that have notepads full of exactly every mouse click and keypress to do their jobs, and heaven help if they miss a step one time--switching from Windows to Apple and wondering how well that'd go.
Please don't bring sex or age into this - I have known people of all ages and genders that do just what you are describing; that is hardly exclusive to middle-aged women.
In fact, it's actually worse with the younger generations who have only grown up with mobile devices and can't understand how to type properly or use anything without a touch interface.

No one here is saying that everyone has to switch from Windows to MacOS, so I'm not sure why you are even bringing that up.
We are talking about the CPU ISA changes and what that will bring to the industry as a whole, and the paradigm shifts it presents, not changing operating systems.
 

UnknownSouljer

Supreme [H]ardness
Joined
Sep 24, 2001
Messages
7,028
I consider Apple's computer hardware and OS market share numbers as niche. A strong niche, but still niche.
Everyone does. If you're not on Windows as a consumer, you're niche.
Technology is cheap to manufacture en masse nowadays...Apple is putting cheap hardware in cheap to make packages (just like every other hardware company) and slapping on some of the most massive price tags while marketing them as trendy home decor and fashion accessories. That's a facepalm that Apple has been eliciting for well over a decade.
Love them or hate them, Apple makes 100% bespoke products. Every PCB, surface, keyboard, mousing surface, audio module is built by them. The only things that are provided by OEM's are things that sit under other people's standards. Such as Bluetooth chips and Wi-Fi radios, and display panels. At this point with ARM, they make, in Apple terminology, the entire widget.
And it turns out when you make bespoke products like this, they're expensive.

As when Microsoft as an example attempted to make a similar product with the Surface, their pricing more or less had to end up where Apple's pricing is. The same thing with their Surface Studio which was an iMac competitor (they haven't even bothered to make a second one as far as I can tell).

Apple doesn't cut corners. They may make decisions that are baffling and it isn't to say they don't make errors (notably Butterfly keyboards for 3 years) - but to say that any of their decisions are "cheap" or even similar to any other OEM, that's just inaccurate. They have the best touch points, highest levels of display quality, and they cram technology into every square inch (to the point that their haptics through their haptic engine are so good, they don't even need a physical button on their all glass touch pad - a multi-microphone system that is good enough to record vocals off of, and sound that is better than a majority of their competitors, etc). They don't stick in lowest common denominator hardware - like TN panels, the cheapest speakers and microphones possible. Every interior space of a Mac is used and has something crammed into it - which is very different than what an OEM PC looks like.

Now, where I will find a place of agreement with you is the premium cost they charge for certain system upgrades such as RAM. It's long been the area that, yeah, it would be much cheaper to just install your own later. And if you buy an iMac or Mac Pro you have that option. But let's be real, Apple isn't the only one charging a pretty hefty premium for RAM and SSD upgrades - every OEM is doing it. But at least when you're in the Apple ecosystem, when it's time to sell you'll likely get your money back from a premium machine when it's time to sell. You won't get that sort of deal on any PC.

Apple didn't become a $1 trillion+ megacorp because of their low low prices.
Marketing and brand recognition, along with high pricing, has worked for them as a "premium" product provider for over 20 years now, so why now would they change the game?

Doing so would only make their market more competitive, which is the opposite of what they want to happen, and even purely from a business standpoint, it would be a foolish move.
Not sure how that is a facepalm, as obviously it is working quite well for them - that facepalm should go to Microsoft for being late to the game, stifling innovation, and dropping the ball for nearly 20 years.
Right, and to add on, they're the biggest company in the world at this point. What's the point in selling more units? It's about making more money. And even on that note they're excelling. The M1 basically performs faster than any PC that costs twice as much. A $900 Mac Mini can destroy a PC that costs $3000 in 4k and h.265 rendering as well as RAW workflows. Especially in real time - and this is just their first generation product. So even going back to the cost and value proposition, Apple is really ahead in specific workloads. Granted again that isn't "gaming", but if that's your (the proverbial you, not you in particular) focus then this whole discussion is pointless.
 
Last edited:

Aurelius

2[H]4U
Joined
Mar 22, 2003
Messages
3,162
Through all of this, I think Apple has recently dropped the ball when they had a golden opportunity to propel ARM. New M1 seems quite stout, as well as significantly cheaper to make and implement than x86. Apple probably should have used those facts to their advantage with a massive advertising smear campaign against PC to educate the masses and lower their base MSRPs, but they didn't. They are still too proud of their pricing, and that's why Apple has only about 8% of the personal computer global hardware market and around 10% of the OS global market share. Windows is around 80%, globally.
I consider Apple's computer hardware and OS market share numbers as niche. A strong niche, but still niche.

Technology is cheap to manufacture en masse nowadays...Apple is putting cheap hardware in cheap to make packages (just like every other hardware company) and slapping on some of the most massive price tags while marketing them as trendy home decor and fashion accessories. That's a facepalm that Apple has been eliciting for well over a decade.
Attack ads aren't really Apple's thing these days, even if it did bring John Hodgman back for a one-off. And while Apple did cut the price of the MacBook Air with the M1 switch, you probably weren't going to see any other price changes with that first round regardless — the initial machines were existing models with new innards. If there are cost savings to be had, they'll show up once Apple is designing for ARM from scratch. Not that I expect to see a $600 MacBook any time soon.

I don't know how much it costs Apple to make an M1, but the hardware isn't cheap. That and a lot of Windows PC makers skew too far in the other direction, where they operate on razor-thin margins and resort to things like poor build quality or bloatware to lower costs. Ideally, PC makers would have thinner profit margins than Apple but still have enough that they don't compromise on design and software.
 
Last edited:
Top