That's not even it. Looking at their 2019 investor presentation, they've been losing PC revenue since 2012. They attribute this to the rise of tablets and smartphones. They own a significant share of many data centers and customers that run ESX have to have a very significant reason to shift to AMD because they'll incur downtime for the VM. AMD has great products but Intel views them as less of a threat than nVidia, who is making big inroads to data centers with AI, the Mellanox acquisition, and general GPU compute. Their big market opportunities are edge compute (Mobileye) and graphics (Xe).
Zen 3 will change that, and turn AMD into a bigger threat. Particularly if Rocket Lake cannot deliver (or can only deliver at insane TDPs as is expected..).
It hurts Intel in that they lost their premier customer in Apple. It's not so much that they lost a volume contract, but this move implies the very same thing that Apple's move from PPC to Intel did--that Intel couldn't cut it and Apple decided that they can do better. When you think about it, that's a pretty big slap at Intel, as they had been the desktop CPU leader for a long time. I'm sure Intel investors aren't happy about this.
Intel probably was going to lose Apple either way. Let’s see what intel does with alder lake, et. al. It’s Curious that some of the technologies on intel’s upcoming roadmap seem to be related to Apple’s needs. For example, alder lake is a hybrid cpu with big and little cores, very similar to the m1. Also, intel is working on upgrading its integrated graphics in tiger lake and alder lake. Apple’s m1 also has upgraded integrated graphics. Could be this is what Apple wanted all along, but intel was too slow to bring it to market.
Uhhhm, nope! Not if Intel was able to surpass their 2016 performance of their i7-6700k and i9-6900k processors, back in 2018, and they were able to pull a meaningful upgrade to r9-5950X levels of performance. It would've meant the Apple Silicon wasn't as competitive, so this decision to shift-microarchitecture would have to remain on the backburner. So this change may never have happened. But to do so would've required Intel to move from their 16nm nodes (+++14nm) nodes, halved-down to 8nm nodes (10nm) successfully. But hitting mass production and retail before 2018 (just before AMD's Threadripper v2). Of course I'm referencing Desktop CPUs here, so just think about those improvements applying to their 15W and 30W Laptop CPUs that we see in the MacBook Air and MacBook Pro.
By the way, this is the entire reason Apple started their side project to sell "The iPad Pro" back in 2016, with a 13 inch screen, removable keyboard, and a more capable iOS system. It was a plan B. Well, now their plan B has turned into plan A.
Alder Lake does not mimic the M1 but ARM's big.LITTLE scheme which ARM developed far earlier than Apple. Alder Lake's success will largely depend on whether Intel's last 10nm iteration (formerly 10nm+++, then rebranded 10nm++ and then I guess, er, SuperFin+?) before 7nm can deliver high enough clocks for its "big" cores at reasonable TDPs.
It's possible that if Intel had delivered the goods they wouldn't have been ditched. On the other hand Apple is all about control, so my guess is that would at best only delay the inevitable. I doubt even Alder Lake's Golden Cove cores will be able to surpass in performance the successor core of Firestorm; I also strongly doubt Alder Lake's small cores will be able to surpass in efficiency even Apple's current Icestorm cores.
Perhaps, even later (in 2024?), Intel's Ocean Cove cores will be able to compete in performance though. I heard they were supposed to be Intel's first truly clean sheet design since at least the original Core processors (though that was way before Jim Keller left Intel, so you never know). I had heard the target IPC was 80% higher than Skylake, which is quite impressive if Intel can pull that. Fabbed at 7nm of course. So more likely 2025+ ;p
dont listen to these guys. its a huge loss. perhaps the biggest intel has suffered in the last 5 years, They just lost 10% of the market lost overnight, and this is the most premium end of the market. ON top of that this is will be biggest road paver for ARM adoption everywhere, devs will use ARM in the cloud if they have ARM boxes. Apple itself is a huge datacenter operator, more software will work properly w. ARM this is just awful news for intel. The ppl say its not a huge deal remind me Obama saying ISIS was the JV team, and yemen is an example of CT success
Also, don’t forget this M1 SOC doesn’t have eGPU support either. This thunderbolt implementation in the new m1 seems like a downgrade as compared to old intel implementation. Also curious is why Apple eschewed thunderbolt4 support...
USB4 support. That's HUGE. They're the first to market & will push the rest of the PC market forward w/ the USB 3.0 nonsense this year & last.
USB4 has Thunderbolt 3 integrated except for mobile. Since Thunderbolt 4 has the same bandwidth as Thunderbolt 3. It's a no-brainer. It helps w/ the longeviity of these devices significantly.
USB4 support is imminent in similar 5nm chips in *ALL* roadmaps publicly revealed by AMD & Intel
"This thunderbolt implementation in the new m1 seems like a downgrade as compared to old intel implementation." - It now has a superset of those capabilities. The only downside is regarding multiple external displays.
"Also curious is why Apple eschewed thunderbolt4 support" - because they'd have to buy it from Intel, and who knows what fresh hell the integration process for that would be. I'm not sure there are any genuine advantages they'd get from implementing Thunderbolt 4.
Apple didn't downgrade their thunderbolt implementation on purpose. The fact is that it takes a while in the design phase to implement new standards. This is one of the first and only ARM CPUs on a mainstream computer, and they managed to implement thunderbolt on it without intel thunderbolt controllers. That is a feat in itself.
Apple probably started designing/implementing the M1 chip long before the new thunderbolt 4/usb 4 spec was announced. This is their first gen, lower powered chip, It doesn't have enough pcie lanes for 4 thunderbolt ports, it can't drive enough displays, this is first gen. It's to be expected.
Intel developed thunderbolt 4, and they are the ones who built it first into their 11th gen chips and simply supplied it to OEMs, it's why it was implemented so fast, if it was optional and OEMs had to do it themselves, buying their own controllers etc....Then thunderbolt 4 wouldn't have been implemented so quick with windows PCs.
The highlight this year is the transition, performance, optimization, Webcam ISP improvement, Neural engine, power efficiency, battery life etc.
They stated, they are building a "Family of Chips". So the one to watch out for is the M2 or M3 for the Macbook Pro 16. That will most definitely have 4 x thunderbolt 4/USB 4 ports....support more displays, have a more powerful GPU.
The M1 has 8 GPU Cores and 2.6 Tflops at 10W......Imagine a 40W Variant with 4 times the GPU cores.....32 cores at 10.4 Tflops.....PS5 level performance in a 40W SOC. It might not scale perfectly linearly, but we know for a fact that a 5nm Apple Silicon 40-60W Macbook Pro 16 Chip will definitely be a beast that will outclass most high end gaming laptops and next (current) gen consoles.
I'm definitely excited for the future of apple soc's and arm in general for more power efficient computing for all major platforms.
What is the strategy for more powerful GPU for game/compute people? They dropped nVidia for AMD over the past few years. Maybe they go back since AMD seems tied to x64 now? Or maybe even Intel?
You think Apple is going to adopt Nvidia in their datacenter because of ARM? They'll just do their own chips if that's what they're wanting to do. I don't think datacenter needs to be ARM based, so they'll likely just go with similar-ish hardware there for now. But if they're really going to be converting their entire lineup to their own chips (including Mac Pro), then that would seem to indicate they're gonna have their own chips for that. Which, my guess is they'll maybe move back to a rack like layout where if you want more CPU cores, they'll slot in extra boards that contain a SoC, and then for various tasks they'll have dedicated boards (like their Afterburner card) and one for AI. I don't think they care about graphics that much, but they could do a larger GPU if needed. But for most of the tasks Apple is concerned with I expect that they'll be able to scale up. I could see them introduce probably a couple of more higher core count chips (I'd expect a 16 to go in the iMacs and higher end MBP, and then maybe one higher than that - I'd guess 24-32, but perhaps they'd go big just for Mac Pro and their own servers).
It's expected the Mac Pro will still support external GPUs for the sake of common sense expandability. No need to lock down the GPU at all. Most Pro users don't rely on integrated graphics.
They may have a M1 variant that doesn't have GPU cores at all unless a buyer wants that as an option.
2 ports on a MBP and it can drive only 1 display? That's no good. If you're connecting headphones + display, you'll have to take off one for charging. Nice.
Headphones still have their own port. Apple basically expect their customers to use wireless headphones, mice etc. anyway though. They'd also probably expect you to charge off the display using Thunderbolt power delivery.
I'm honestly not sure that would happen. If you check the performance analysis of A14, Zen 3 (on the 5950X) is superior on a single-core basis, but only by a little. Zen 2 would be some way behind.
"The SoC supports an odd combination of Thunderbolt 3 and the new USB 4 (but not the all-encompassing Thunderbolt 4)"
How is this odd? Isn't TB4 still exclusive to Intel, with TB3 being integrated into the USB 4 spec? As such, isn't this the only configuration one could reasonably expect?
The looming issue is software. Unlike the PPC -> x86 transition (where PPC emulation on x86 was rarely slower and often faster than native PPC anyway) x86 on ARM is... not good. That makes Apple reliant on developers re-writing their software for ARM in order for their new devices not to take a performance nosedive in day-to-day usage. On top of that, many of those developers already HAVE an ARM codebase: the one they use for iOS. Now imagine you are a Mac developer, Apple has given you just a year to transition your codebase to a different CPU architecture, and you already have a codebase on that new architecture. Do you a) rewrite your massive desktop focused application stack within a year, or b) tart up your existing ARM codebase with a desktop UI, and maybe think about rewriting your desktop application at some point in the future (read: no).
"x86 on ARM is... not good" As far as I'm aware, our only data point for that is based on Windows running on what is effectively an overclocked Snapdragon 855.
It's notable which devices they've chosen to replace first, though - these were all running on outdated Intel chips that struggled for IPC and couldn't hit high boost clocks under heavy load. I suspect any deficit in performance from code translation will be somewhat compensated by better behaviour of the SoC in these particular devices relative to their predecessors.
By the time they start replacing higher-end devices, we should have a better picture of real-world performance.
What befuddles me is the lack of LTE/5G modem. It seems apparent that market has value owing to the LTE variants of the iPad and the Qualcomm "always connected" 8cx strategy (even though that SoC is off to a rough start).
@ Ryan Smith: I noticed you listed all of the Apple M1 configurations as having LPDDR4X-4266 memory - do we have confirmation on this? I thought last I heard, it was rumored that the A14 silicon possibly had made the switch to LPDDR5; in which case, of course M1 would also be using LPDDR5.
Thanks Ryan! These models are a bit disappointing, especially the Pro series. Starting with the same old 720p webcam that just won't go away (even low-end new iPhones have a far better selfie cam), continuing on to the external connectivity (two ports on the Pro, that's it) and the inability to use eGPUs, including the ones sold by Apple itself. And, there is the price: very Apple. That M1 SoC is a major attraction, the rest awaits an overhaul.
It's only disappointing if you forget what the target market for the M1 is. In their presentation, right up front, Apple said it's optimized for battery life. So, it's their version of Intel's U-series chips - low tdp, low-ish performance, for ultraportables, and not the high performance you're wanting if you are a developer.
Hold on to your hat in a few months when the M2 (or M1X) rolls out for the high end MBP models - which will have a higher TDP, support for more RAM, more ports, and be much, much faster.
My tea leaves tell me there's probably going to be an M3 as well, with desktop-level TDP and support for discrete graphics.
The fact that apple's lowest-end, lowest powered chip is boasting such incredible speed increases suggests that we haven't seen anything yet.
Regarding the history of Apple wanting its own silicon: a good while ago, a former (!) Intel engineer stated that Apple basically "had it" with Intel when Intel just couldn't get the bugs in its CPUs ironed out a few generations ago. The inability to move to more efficient 10 and 7 nm nodes didn't help, either. Intel took the eye of the fab technology ball some years ago, and it has bitten them in the rear ever since. They are still a power house, but playing catch-up is hard. One indicator that would be positive is if and when Intel buys EUV hardware in some numbers; I don't believe that has even started.
Intel kept arriving late with the 28W chips and/or not executing on the GPU front properly, too. The writing's been on the wall for a long time - Charlie Demerjian called it ages ago, just been waiting for the world to catch up 😅
Mobile Campus Engadine provides the best iPhone Screen replacement, mobile phone repairs, Battery Replacement, Water Damage Repair in NSW Sydney. Get a Free Quote Now. https://g.page/mobile-campus-engadine
Mobile phones, tablets, iPads and related accessories is now a basic necessity of everyone’s life. Alfalah Mobiles offers a variety of repair services to the customers of its vicinity. We are located at the heart of Punchbowl market, near to Punchbowl train station and Woolworths. We offer mobile phone repair services with specialty in iPhone screen replacement and Samsung phone screen replacement. Don't hesitate to give us a call or visit to fix your phone battery or camera problems. We never disappoint our customers in terms of cheap rates and quality of services. We have cheap, durable, fancy and the latest technology accessories for your mobile phones, regardless how old or new your phone is, including iPhone X and Samsung Galaxy S10. https://alfalahmobiles.com.au/
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
49 Comments
Back to Article
Zandros - Tuesday, November 10, 2020 - link
As far as I can tell, there's unfortunately no 10 Gb Ethernet available for the Mac mini.Ryan Smith - Wednesday, November 11, 2020 - link
Hmm. I swear I saw it earlier when double-checking the tech specs. But right you are: it's not listed now. Thanks for the heads up!Slash3 - Wednesday, November 11, 2020 - link
HDMI is also missing from the feature chart for the 2020 model.melgross - Wednesday, November 11, 2020 - link
There’s an HDMI 2 port on the back. Look at the pic.shabby - Tuesday, November 10, 2020 - link
How many laptops does apple sell, is this going to hurt intel at all?goatfajitas - Tuesday, November 10, 2020 - link
Not a ton. Apple is just one vendor... but Intel is already hurting, milking 14nm cause for 6 years and being surpassed by AMD are far bigger issuesgoatfajitas - Tuesday, November 10, 2020 - link
Cause = stupid spellchecker CPU'sHappyCracker - Wednesday, November 11, 2020 - link
That's not even it. Looking at their 2019 investor presentation, they've been losing PC revenue since 2012. They attribute this to the rise of tablets and smartphones. They own a significant share of many data centers and customers that run ESX have to have a very significant reason to shift to AMD because they'll incur downtime for the VM. AMD has great products but Intel views them as less of a threat than nVidia, who is making big inroads to data centers with AI, the Mellanox acquisition, and general GPU compute. Their big market opportunities are edge compute (Mobileye) and graphics (Xe).Santoval - Wednesday, November 11, 2020 - link
Zen 3 will change that, and turn AMD into a bigger threat. Particularly if Rocket Lake cannot deliver (or can only deliver at insane TDPs as is expected..).Santoval - Wednesday, November 11, 2020 - link
Soon to be 7 years (when Rocket Lake is released) and counting...melgross - Wednesday, November 11, 2020 - link
About 15 million a year. Last quarter saw a large jump though.MonkeyPaw - Tuesday, November 10, 2020 - link
It hurts Intel in that they lost their premier customer in Apple. It's not so much that they lost a volume contract, but this move implies the very same thing that Apple's move from PPC to Intel did--that Intel couldn't cut it and Apple decided that they can do better. When you think about it, that's a pretty big slap at Intel, as they had been the desktop CPU leader for a long time. I'm sure Intel investors aren't happy about this.Dehjomz - Tuesday, November 10, 2020 - link
Intel probably was going to lose Apple either way. Let’s see what intel does with alder lake, et. al. It’s Curious that some of the technologies on intel’s upcoming roadmap seem to be related to Apple’s needs. For example, alder lake is a hybrid cpu with big and little cores, very similar to the m1. Also, intel is working on upgrading its integrated graphics in tiger lake and alder lake. Apple’s m1 also has upgraded integrated graphics. Could be this is what Apple wanted all along, but intel was too slow to bring it to market.Kangal - Wednesday, November 11, 2020 - link
Uhhhm, nope!Not if Intel was able to surpass their 2016 performance of their i7-6700k and i9-6900k processors, back in 2018, and they were able to pull a meaningful upgrade to r9-5950X levels of performance. It would've meant the Apple Silicon wasn't as competitive, so this decision to shift-microarchitecture would have to remain on the backburner. So this change may never have happened. But to do so would've required Intel to move from their 16nm nodes (+++14nm) nodes, halved-down to 8nm nodes (10nm) successfully. But hitting mass production and retail before 2018 (just before AMD's Threadripper v2). Of course I'm referencing Desktop CPUs here, so just think about those improvements applying to their 15W and 30W Laptop CPUs that we see in the MacBook Air and MacBook Pro.
By the way, this is the entire reason Apple started their side project to sell "The iPad Pro" back in 2016, with a 13 inch screen, removable keyboard, and a more capable iOS system. It was a plan B. Well, now their plan B has turned into plan A.
Santoval - Wednesday, November 11, 2020 - link
Alder Lake does not mimic the M1 but ARM's big.LITTLE scheme which ARM developed far earlier than Apple. Alder Lake's success will largely depend on whether Intel's last 10nm iteration (formerly 10nm+++, then rebranded 10nm++ and then I guess, er, SuperFin+?) before 7nm can deliver high enough clocks for its "big" cores at reasonable TDPs.It's possible that if Intel had delivered the goods they wouldn't have been ditched. On the other hand Apple is all about control, so my guess is that would at best only delay the inevitable. I doubt even Alder Lake's Golden Cove cores will be able to surpass in performance the successor core of Firestorm; I also strongly doubt Alder Lake's small cores will be able to surpass in efficiency even Apple's current Icestorm cores.
Perhaps, even later (in 2024?), Intel's Ocean Cove cores will be able to compete in performance though. I heard they were supposed to be Intel's first truly clean sheet design since at least the original Core processors (though that was way before Jim Keller left Intel, so you never know). I had heard the target IPC was 80% higher than Skylake, which is quite impressive if Intel can pull that. Fabbed at 7nm of course. So more likely 2025+ ;p
azfacea - Wednesday, November 11, 2020 - link
dont listen to these guys. its a huge loss. perhaps the biggest intel has suffered in the last 5 years, They just lost 10% of the market lost overnight, and this is the most premium end of the market. ON top of that this is will be biggest road paver for ARM adoption everywhere, devs will use ARM in the cloud if they have ARM boxes. Apple itself is a huge datacenter operator, more software will work properly w. ARM this is just awful news for intel. The ppl say its not a huge deal remind me Obama saying ISIS was the JV team, and yemen is an example of CT success0iron - Wednesday, November 11, 2020 - link
The transition period is two years, not overnight. I'm not sure what % of those 3 models accounts in total of Mac shipment.ingwe - Wednesday, November 11, 2020 - link
Well two years for a change of this size is pretty much overnight.Dehjomz - Tuesday, November 10, 2020 - link
Also, don’t forget this M1 SOC doesn’t have eGPU support either. This thunderbolt implementation in the new m1 seems like a downgrade as compared to old intel implementation. Also curious is why Apple eschewed thunderbolt4 support...Dolda2000 - Tuesday, November 10, 2020 - link
>doesn’t have eGPU support eitherHow do you know? I haven't seen that mentioned anywhere.
nyoungman - Tuesday, November 10, 2020 - link
Here's an article comparing the "Other Accessories":https://www.macrumors.com/2020/11/10/m1-macs-not-c...
ingwe - Wednesday, November 11, 2020 - link
Thanks for sharing the link. That is a big disappointment.Spunjji - Thursday, November 12, 2020 - link
Makes sense. They'd have to get an entirely new driver stack written, and eGPU on Mac is a niche within a niche.lilkwarrior - Wednesday, November 11, 2020 - link
USB4 support. That's HUGE. They're the first to market & will push the rest of the PC market forward w/ the USB 3.0 nonsense this year & last.USB4 has Thunderbolt 3 integrated except for mobile. Since Thunderbolt 4 has the same bandwidth as Thunderbolt 3. It's a no-brainer. It helps w/ the longeviity of these devices significantly.
USB4 support is imminent in similar 5nm chips in *ALL* roadmaps publicly revealed by AMD & Intel
Spunjji - Thursday, November 12, 2020 - link
"This thunderbolt implementation in the new m1 seems like a downgrade as compared to old intel implementation." - It now has a superset of those capabilities. The only downside is regarding multiple external displays."Also curious is why Apple eschewed thunderbolt4 support" - because they'd have to buy it from Intel, and who knows what fresh hell the integration process for that would be. I'm not sure there are any genuine advantages they'd get from implementing Thunderbolt 4.
kninez - Friday, November 13, 2020 - link
Apple didn't downgrade their thunderbolt implementation on purpose. The fact is that it takes a while in the design phase to implement new standards. This is one of the first and only ARM CPUs on a mainstream computer, and they managed to implement thunderbolt on it without intel thunderbolt controllers. That is a feat in itself.Apple probably started designing/implementing the M1 chip long before the new thunderbolt 4/usb 4 spec was announced. This is their first gen, lower powered chip, It doesn't have enough pcie lanes for 4 thunderbolt ports, it can't drive enough displays, this is first gen. It's to be expected.
Intel developed thunderbolt 4, and they are the ones who built it first into their 11th gen chips and simply supplied it to OEMs, it's why it was implemented so fast, if it was optional and OEMs had to do it themselves, buying their own controllers etc....Then thunderbolt 4 wouldn't have been implemented so quick with windows PCs.
The highlight this year is the transition, performance, optimization, Webcam ISP improvement, Neural engine, power efficiency, battery life etc.
They stated, they are building a "Family of Chips". So the one to watch out for is the M2 or M3 for the Macbook Pro 16. That will most definitely have 4 x thunderbolt 4/USB 4 ports....support more displays, have a more powerful GPU.
The M1 has 8 GPU Cores and 2.6 Tflops at 10W......Imagine a 40W Variant with 4 times the GPU cores.....32 cores at 10.4 Tflops.....PS5 level performance in a 40W SOC. It might not scale perfectly linearly, but we know for a fact that a 5nm Apple Silicon 40-60W Macbook Pro 16 Chip will definitely be a beast that will outclass most high end gaming laptops and next (current) gen consoles.
I'm definitely excited for the future of apple soc's and arm in general for more power efficient computing for all major platforms.
Pinn - Tuesday, November 10, 2020 - link
What is the strategy for more powerful GPU for game/compute people? They dropped nVidia for AMD over the past few years. Maybe they go back since AMD seems tied to x64 now? Or maybe even Intel?darkswordsman17 - Wednesday, November 11, 2020 - link
You think Apple is going to adopt Nvidia in their datacenter because of ARM? They'll just do their own chips if that's what they're wanting to do. I don't think datacenter needs to be ARM based, so they'll likely just go with similar-ish hardware there for now. But if they're really going to be converting their entire lineup to their own chips (including Mac Pro), then that would seem to indicate they're gonna have their own chips for that. Which, my guess is they'll maybe move back to a rack like layout where if you want more CPU cores, they'll slot in extra boards that contain a SoC, and then for various tasks they'll have dedicated boards (like their Afterburner card) and one for AI. I don't think they care about graphics that much, but they could do a larger GPU if needed. But for most of the tasks Apple is concerned with I expect that they'll be able to scale up. I could see them introduce probably a couple of more higher core count chips (I'd expect a 16 to go in the iMacs and higher end MBP, and then maybe one higher than that - I'd guess 24-32, but perhaps they'd go big just for Mac Pro and their own servers).lilkwarrior - Wednesday, November 11, 2020 - link
It's expected the Mac Pro will still support external GPUs for the sake of common sense expandability. No need to lock down the GPU at all. Most Pro users don't rely on integrated graphics.They may have a M1 variant that doesn't have GPU cores at all unless a buyer wants that as an option.
lilkwarrior - Wednesday, November 11, 2020 - link
It'd be interesting if Apple started created their own discrete GPUs. That would certainly get AMD & Nvidia's attentionTeckk - Wednesday, November 11, 2020 - link
2 ports on a MBP and it can drive only 1 display? That's no good. If you're connecting headphones + display, you'll have to take off one for charging. Nice.Spunjji - Thursday, November 12, 2020 - link
Headphones still have their own port. Apple basically expect their customers to use wireless headphones, mice etc. anyway though. They'd also probably expect you to charge off the display using Thunderbolt power delivery.AshlayW - Wednesday, November 11, 2020 - link
Excited for real world performance analysis. Be interesting to see it get slapped around by Renoir.Spunjji - Thursday, November 12, 2020 - link
I'm honestly not sure that would happen. If you check the performance analysis of A14, Zen 3 (on the 5950X) is superior on a single-core basis, but only by a little. Zen 2 would be some way behind.Valantar - Wednesday, November 11, 2020 - link
"The SoC supports an odd combination of Thunderbolt 3 and the new USB 4 (but not the all-encompassing Thunderbolt 4)"How is this odd? Isn't TB4 still exclusive to Intel, with TB3 being integrated into the USB 4 spec? As such, isn't this the only configuration one could reasonably expect?
Vanburen - Wednesday, November 11, 2020 - link
Would there be any benefits to moving to HBM instead of LPDDR4X for the M1?edzieba - Wednesday, November 11, 2020 - link
The looming issue is software. Unlike the PPC -> x86 transition (where PPC emulation on x86 was rarely slower and often faster than native PPC anyway) x86 on ARM is... not good. That makes Apple reliant on developers re-writing their software for ARM in order for their new devices not to take a performance nosedive in day-to-day usage. On top of that, many of those developers already HAVE an ARM codebase: the one they use for iOS.Now imagine you are a Mac developer, Apple has given you just a year to transition your codebase to a different CPU architecture, and you already have a codebase on that new architecture. Do you a) rewrite your massive desktop focused application stack within a year, or b) tart up your existing ARM codebase with a desktop UI, and maybe think about rewriting your desktop application at some point in the future (read: no).
lilkwarrior - Wednesday, November 11, 2020 - link
For that reason, all iPhone apps are now usuable on the new Macs towards a universal app approach.Spunjji - Thursday, November 12, 2020 - link
"x86 on ARM is... not good"As far as I'm aware, our only data point for that is based on Windows running on what is effectively an overclocked Snapdragon 855.
It's notable which devices they've chosen to replace first, though - these were all running on outdated Intel chips that struggled for IPC and couldn't hit high boost clocks under heavy load. I suspect any deficit in performance from code translation will be somewhat compensated by better behaviour of the SoC in these particular devices relative to their predecessors.
By the time they start replacing higher-end devices, we should have a better picture of real-world performance.
twnznz - Wednesday, November 11, 2020 - link
What befuddles me is the lack of LTE/5G modem. It seems apparent that market has value owing to the LTE variants of the iPad and the Qualcomm "always connected" 8cx strategy (even though that SoC is off to a rough start).NextGen_Gamer - Wednesday, November 11, 2020 - link
@ Ryan Smith: I noticed you listed all of the Apple M1 configurations as having LPDDR4X-4266 memory - do we have confirmation on this? I thought last I heard, it was rumored that the A14 silicon possibly had made the switch to LPDDR5; in which case, of course M1 would also be using LPDDR5.eastcoast_pete - Wednesday, November 11, 2020 - link
Thanks Ryan! These models are a bit disappointing, especially the Pro series. Starting with the same old 720p webcam that just won't go away (even low-end new iPhones have a far better selfie cam), continuing on to the external connectivity (two ports on the Pro, that's it) and the inability to use eGPUs, including the ones sold by Apple itself. And, there is the price: very Apple. That M1 SoC is a major attraction, the rest awaits an overhaul.Glaurung - Wednesday, November 11, 2020 - link
It's only disappointing if you forget what the target market for the M1 is. In their presentation, right up front, Apple said it's optimized for battery life. So, it's their version of Intel's U-series chips - low tdp, low-ish performance, for ultraportables, and not the high performance you're wanting if you are a developer.Hold on to your hat in a few months when the M2 (or M1X) rolls out for the high end MBP models - which will have a higher TDP, support for more RAM, more ports, and be much, much faster.
My tea leaves tell me there's probably going to be an M3 as well, with desktop-level TDP and support for discrete graphics.
The fact that apple's lowest-end, lowest powered chip is boasting such incredible speed increases suggests that we haven't seen anything yet.
eastcoast_pete - Wednesday, November 11, 2020 - link
Regarding the history of Apple wanting its own silicon: a good while ago, a former (!) Intel engineer stated that Apple basically "had it" with Intel when Intel just couldn't get the bugs in its CPUs ironed out a few generations ago. The inability to move to more efficient 10 and 7 nm nodes didn't help, either.Intel took the eye of the fab technology ball some years ago, and it has bitten them in the rear ever since. They are still a power house, but playing catch-up is hard. One indicator that would be positive is if and when Intel buys EUV hardware in some numbers; I don't believe that has even started.
incanter - Wednesday, November 11, 2020 - link
> Regarding the history of Apple wanting its own siliconIt might be worth mentioning that Power PC was (33%) Apple's silicon... for as much good as it did them.
Spunjji - Thursday, November 12, 2020 - link
Intel kept arriving late with the 28W chips and/or not executing on the GPU front properly, too. The writing's been on the wall for a long time - Charlie Demerjian called it ages ago, just been waiting for the world to catch up 😅peevee - Friday, November 13, 2020 - link
Prices for individual spec configurations instead of just the minimum would be greatly appreciated.Also, WiFi specs, other than MBP, are not given.
mobilecampus - Monday, September 6, 2021 - link
Mobile Campus Engadine provides the best iPhone Screen replacement, mobile phone repairs, Battery Replacement, Water Damage Repair in NSW Sydney. Get a Free Quote Now.https://g.page/mobile-campus-engadine
mobilecampus - Wednesday, October 6, 2021 - link
Mobile phones, tablets, iPads and related accessories is now a basic necessity of everyone’s life.Alfalah Mobiles offers a variety of repair services to the customers of its vicinity.
We are located at the heart of Punchbowl market, near to Punchbowl train station and Woolworths.
We offer mobile phone repair services with specialty in iPhone screen replacement and Samsung phone screen replacement.
Don't hesitate to give us a call or visit to fix your phone battery or camera problems.
We never disappoint our customers in terms of cheap rates and quality of services.
We have cheap, durable, fancy and the latest technology accessories for your mobile phones, regardless how old or new your phone is, including iPhone X and Samsung Galaxy S10.
https://alfalahmobiles.com.au/