Meh. Who cares about speed when most of my apps won't run in MacOS
Thanks to the Rosetta 2 translation layer, they pretty much all run (there are a few minor glitches so far). What apps specifically do you use that you think will not run?
It's called Corsair iCUE on macOS and Logitech G Hub is also available too. They might not be ported yet but they will be to avoid customer complaints. Your only missing software is video games, which will always be targeted for DirectX and Vulkan/OpenGL graphics on an x86 architecture anyway. At the end of the day, if you need backwards compatibility and long term support, you don't want macOS and everyone knows that.
However, with all that said, macOS is currently the most secure desktop OS by default through isolating GUI apps (like what Wayland aims to achieve on Linux) and applying a sandbox to almost all apps which acts similarly to apparmor+seccomp-bpf but with a few creature comfort helper tools. It even has basic mandatory access controls to stop apps snooping in private user areas and will soon be the only system that has a truly robust Secure Boot implementation alongside a tamper-proof OS partition (a bit like implementing BSD Securelevels and making/usr immutable but far better).
Honestly, the security part of my brain has a wet dream over some of what Apple has seamlessly added to mainstream desktop computing but ultimately the dominant part of my brain which wants 'one computer to rule them all' will always opt for Windows. All I'm missing is proper telephony+SMS+4G baseband support and my desktop PC really can do everything with no need for any of this smartphone, tablet or laptop poop.
The first 3 aren't worth running to start with. Waterfox has a Mac version. I think Logitech and Winamp actually has Mac versions too.
The only valid items on this list is your Steam library. Mac laptops were never suitable for gaming to begin with.
At the same time, since these Macs will be able to support adapted iPad and iPhone games, it may get more gamer friendly. It still won't be "PC master race" when it comes to games, but I am sure most hardcore gamers probably have a custom speced computer anyhow.
The only valid items on this list is your Steam library. Mac laptops were never suitable for gaming to begin with.
And yet, some of them will probably run with WINE, and all will run with Boot Camp or in a virtualization environment. As a non-gamer, it's not a big deal to me, but for some people, it might be.
Mac laptops were never suitable for gaming to begin with.
I've gotten quite a bit of mileage out of my steam library on a 50gb bootcamp partition using a lowly 2018 macbook air that I picked up at launch. I use a desktop for the latest and greatest, of course, but my air serves me well for both work and leisure. I'll probably just move to a PC laptop and run linux with a windows partition next work laptop purchase.
Here's my understanding. Yes, WINE runs on ARM, but it does not do any kind of processor emulation. That means if you take your Windows source code and compile it to target ARM, your binary should work with WINE running on an ARM computer.
Now it's been years since I've used MacOS much, but I do remember the first version of Rosetta, which allowed the old PowerPC binaries to run on new Intel macs. When you opened an app for the first time it would transcompile the PowerPC object code to Intel object code
What apps specifically do you use that you think will not run?
The question you should be asking after all the launch problems with Apple's certificate approval server failing, is what apps in the future will Apple let me run?
The thing that got me originally buying macs was that they ran intel so if I found that there was an app I needed I could boot it into Linux or even Windows. With the new M1 chip all you have is MacOS and nothing else. So if you need to use Linux or Windows for anything that's it, game over you have to buy a PC. The huge price increases, remov
A list of his apps that won't run on MacOS and that he can't find an alternative for would be more interesting than unsubstantiated claims. Especially in view of the fact that the vast majority of FOSS will compile and run on macOS. It used to be the rule around here that if you can't support your claims you were assumed to be full of shit until you did.
There have been comparisons for a while now, that showed modern day iPads were performing as well as, or better than a number of laptops...
So talking about a similar chip but with performance scaled up to a laptop with active cooling (in the case of the MacBook Pro), and to me it's not a surprise that people are seeing the performance Apple claimed they would.
Some of the people I follow on social have been showing equally impressive results on tasks across the board, weather it be hight end 4k video encoding, or compilation tasks for large projects.
For anyone that works day to day on a Mac it is VERY tempting to get even just the Mac mini for a performance boost, even with the 16GB limitation. Will probably wait a while before I take the plunge myself but it's great to see that we are at the dawn of a really promising line of systems that might actually deliver a significant increase of performance for real-world tasks, unlike the last few years of intel updates where often it was indiscernible if anything was really faster at all.
Even in Geekbench, which seems to be tuned to favour Apple parts, the M1 is lagging behind Intel Macbooks. And Macbooks are not fast compared to laptops with better cooling, and certainly not compared to AMD laptops.
So while it's surprisingly decent the fact is it takes longer to export a video in Final Cut on an M1 than it does on an older Intel Macbook Pro. The GPU is similar to mid range AMD and Nvidia mobile chips from 3-4 years ago.
Are we looking at the same Geekbench scores? I see the M1 using x86 emulation beating the native Intel Macbooks, and pulling out far ahead using the native Geekbench.
Still pushing that bullshit, jcr? Rosetta is an emulator, it cannot be anything other than an emulator.
Translation and emulation are orthagonal, not mutually exclusive. You've been told that countless times, but you continue to repeat the same garbage because you're incapable of learning anything.
Interpreter, jcr, what Rosetta isn't is an interpreter. It is an emulator because it emulates an x86 Mac. Anyone technically competent would understand this, but not you.
Right, and in Cinebench we see the Rosetta scores generally matching the previously used Intel chips, and the native scores significantly surpassing them. Yes, they're not as good as Zen 3, but that's not relevant since Apple has never shipped a Zen 3 CPU. The two questions here are, are the M1 macs an improvement over their predecessors, and how much of a performance penalty does the lack of native apps have. And on the former, the answer is "a rather large improvement", and on the latter, the answer is "2
It did on the very first released benchmark results, and never again after that.
A possible explanation: Macs do a lot of work when they are started for the first time (like collecting information so that "search" works instantly), so if you start a brand new Mac, download a benchmark, launch the benchmark _immediately_, it will run slower because your Mac does other things.
With all newer M1 Macs, the order is MacMini consistently very slightly faster than MacBook Pro, and MacBook Air equal to MacBook
Those comparisons of iPads to laptops were shown to be worthless though when the whole 20 minute WildLife benchmark came out a little while ago though. The A series chips after 3-4 minutes just tanked in performance to about 60%, and continued to drop until about 50%ish performance. While this might be fine on a consumption device like a iPad, they aren't on a laptop that is expected to work for more then a few moments at a time.
Its also one of those things that I'm noticing is lacking in these reviews. What
How long does it take to run? Does CPU 2017 take longer than CPU 2006?
Run time depends on the system, suite, compiler, tuning, and how many copies or threads are chosen.
So according to the link you gave me, it will depend on how they ran the test. It says that if you run each test alone, they can be done in a few minutes.
Can you please show/quote where anandtech states how they ran the benchmarking? Was it a full suite that will take hours, or a single test that is done in a few minutes and then lets the system cool down? I tried looking myself, but can't find any indication how the tests were run
So according to the link you gave me, it will depend on how they ran the test. It says that if you run each test alone, they can be done in a few minutes.
Can you please show/quote where anandtech states how they ran the benchmarking? Was it a full suite that will take hours, or a single test that is done in a few minutes and then lets the system cool down? I tried looking myself, but can't find any indication how the tests were run
If you believe you have found such a fundamental flaw in Anandtech's testing, then you might want to inform them of your concerns. I'm sure they would be pleased to have their errors pointed out so they can correct them.
But once again those are artificial benchmarks. Where are the benchmarks of actual applications that people use? The only one in your linked article is Tomb Raider, all the rest are artificial. They didn't even try Cinebench.
Rise of the Tomb Raider, enthusiast M1 - 39 FPS Low end Ryzen 4600H with 1650 - 58 FPS High end i7 10th gen with an old 1660Ti - 82 FPS
Unfortunately they don't give 95 and 99th percentile scores.
For anyone that works day to day on a Mac it is VERY tempting to get even just the Mac mini for a performance boost, even with the 16GB limitation. Will probably wait a while before I take the plunge myself but it's great to see that we are at the dawn of a really promising line of systems that might actually deliver a significant increase of performance for real-world tasks, unlike the last few years of intel updates where often it was indiscernible if anything was really faster at all
While I am tempted by the speed and the novelty of the new, the one thing missing for me is being able to run intel based VMs. And I believe that Apple has said that Rosetta 2 doesn't do that.
I'm almost to the point of replacing my 2014 mac mini with an intel 2018 model, rather than new shiny. And I'm sure a lot of people are in the same boat.
As long as those "real-world tasks" don't require much memory or scaling to multiple cores or x86 OS compatibility, otherwise known as "real world tasks that suit a low power portable processor". But sure, cherry pick the results, as you always do SuperKendall.
But it's the ARM version of Docker (running ARM containers images & ARM app binaries), so probably not helpful unless you have ARM versions of your programs or ability to build them from source.
I doubt it. For one thing this is a Mac processor not a iPhone/iPad processor so mobile apps aren't really an issue in this context:
Apple has introduced the new M1 chip that will power its new generation of Arm-based Macs
Secondly, most of the apps worth using are already available on Android and iOS since iOS users are more profitable than Android users. Finally pretty much all FOSS is compilable on MacOS and iOS and even most of the Microsoft products are available on MacOS these days. I've been using L
Here is a cost based analysis of building ARM-chips themselves versus buying from Intel. Net-net, they save $2B + (and this analysis does not include the RAM costs)
https://medium.com/@sumitg_168... [medium.com]
Non upgradable RAM, mid range performance for high end prices...
Apple, as we all know, does "value" based pricing, not cost-up. And value = brand, apps, ease of use, etc. They are not optimizing for folks who want upgradable RAM, etc. From apple's perspective, those folks can buy the high-end MacBooks... And $40B in revenue per year suggests their product strategy is working.
You are 100% right about value-based pricing, but slashdot readers have never understood and will never understand that. This is a crowd that thinks computers should be priced be weighing the silicon, metal, and plastic that goes into their production and charging a flat rate plus 5% markup.
The money they don't have to contribute to Intel's net profit isn't even the main reason for the transition. By going to their own CPUs, Apple's product release cycles aren't subject to a third party's schedule.
That's hardly surprising. Intel has been riding a gravy train for 40 years, charging $100-$900 for a sliver of silicon that other CPU/SoC vendors could only get $5-$25 for at the same size. The CPU industry was badly in need of competition, which was difficult because what gave Intel the performance lead was its monopoly over proprietary process fabrication plants. They were consistently 1-2 generations ahead of anyone else, which meant they could manufacture faster at the same power draw, or lower power
If you hate Apple enough, it's easy to dismiss what's happening right in front of us at the moment. What I find most interesting is looking at what this M chip can do already but also to keep in mind that they have yet to release their "Pro" spec chip (P-series?).
This is just the first iteration of a mobile level chip and it's already outperforming x86 CPUs and discrete GPUs, albeit older ones. Going the route of previously unseen levels vertical integration also opens up all sorts of new systems development avenues in areas that are up until now unexplored by PC makers.
Given the performance levels we're already seeing, I believe we're all in for quite the shock when they release their P-series chip.
Update: It looks like the Bits and Chips CInebench R23 result is not for the Apple M1 SoC but from an A12Z. MacRumours has a result in that it reckons is from a genuine new M1 13-inch MacBook Pro, though, and it is somewhat better than the A12Z, of course. It quotes Twitter user @mnloona48_ who shared an MBP M1 unboxing and various other tests on his new 8GB machine. Mnloona48's Cinebench R23 scores are an impressive 1,498 single core, 7,508 multi core.
Now I'm impressed. My year old Linux box with a Ryzen 3400G is 1068 single and 4810 multi. And it is entirely fast enough for my use.
The remaining problem with the mini is the lack of ports. I have four usb things plugged in now, and I would need two more to make up for the second drive and the DVD drive. And I still need a plug for a flash drive, so I'd need 8 USB ports on the mini.
The link I posted compared with a mid range 4800 with 35W TDP.
The article at the link you posted has been updated [hexus.net], the M1 is better at single thread and almost the same in multi-threaded performance... on Apple's lowest end laptop. Compared to both mobile and desktop CPU's from AMD.
Unless less points are better on this benchmark, or you think 10,600 is less than 4,530... Your link shows the Ryzen 4800 is a fair bit ahead of the M1. Although I looked up the same benchmark (Cinemark R23) on a different site and it showed them nearly tied.
No idea what's going on there, but my Ryzen 48000H system cost me $1100 and it came with a discrete, gameable GPU (RX 5600M) whereas the $1000 Apple does not. I've had it for 6 months, the price has already dropped since then. This month AMD released t
It is actually quite easy to predict. Expect single core performance to be no more then 5% faster then the M1. Better cooling will facilitate a slightly faster clock rate but the core should be the same. For multi-threaded performance expect ~1.75 times faster with 8 fast cores or ~3.3 times faster for 16 fast cores.
The unknown is the memory architecture, GPU, and PCIe lanes. The internal GPU might be excluded in favor of more CPU cores. The number of memory lanes is unknown but it is possible we se
Agreed. I think the P-series is going to be a game changer for the desktop folks who can really use it.
Right now my concern is the complete lack of virtualization on the platform. The best you can do is an arm-based linux distro. Both Parallels and VMware have reported that they are working on a version of their platforms that will work with the Apple silicon, but I can't imagine that this isn't going to suck as there would have to be a lot of instruction set translation going on.
This is just the first iteration of a mobile level chip and it's already outperforming x86 CPUs and discrete GPUs, albeit older ones.
You have a short memory. Apple used to use PowerPC before switching to Intel. They were RISC-based and also claimed to be much more powerful than Intel CPUs...and yet Macs had a terrible market share. There was a reason that Macs took off when they switched to Intel.
Apple isn't shifting any paradigms with it's m1 processor. This is just TSMC doing what they do. It's the same way with the new AMD processors that TSMC also makes and which beat the older intel stuff by an even larger margin.
As in most of, well anyting that is complex and new technology, I'll allow others to buy the first ones and work out the kinks.
Simple rule for me: Never buy the first generation of technology. I've worked the back side of technology and see how they test things, or just don't...
I think, for a lot of people, there's just not a compelling case to be on the leading edge with new tech anymore. We've been at a point for a number of years where the performance of even relatively "old" tech outstrips what most people need.
I can't say that I ever feel constrained writing code on this 2015 MacBook Pro, for instance.
by Anonymous Coward writes:
on Tuesday November 17, 2020 @06:59PM (#60735876)
All the comments compare the M1 to the writer's own "expectations", which is meaningless. Actual objective benchmarks show it's nowhere near a decent gaming / workstation PC. Couple that with Apple's walled garden, anti-repair and anti-recycling stance, and terrible ethics, and it's a very solid "no".
I hope that Apple's success with ARM will spur the rest of the ARM industry to get its act together and get ARM standardized into a complete and compatible platform that Linux and Windows can work with. Right now the ARM landscape outside of this new Apple ecosystem is quite a wasteland full of incompatible and proprietary boot systems and requiring custom kernels, device drivers, hardware trees, etc.
When we get to the point that a person can buy an ARM laptop, server, or desktop mini PC, and then go to RedHat.com and get stock RHEL for it, or go to Fedora, Ubuntu, or Arch, and get a stock distribution for it, or even a Windows 10 ISO from Microsoft, then Intel start to find themselves in trouble. Right now they have little to worry about in the PC world, other than AMD. Except for some hobby or embedded applications, the way things are currently, I would pick an Intel or AMD CPU any day over ARM. And that includes my router firewall. I want that to change!
It's not that ARM can't perform. It's that each ARM system is a world of its own right now. I've heard about some standards being proposed and attempted, but I've not seen much evidence of movement there. Currently most vendors are satisfied to supply binary blobs for proprietary Android installations.
Right now the ARM landscape outside of this new Apple ecosystem is quite a wasteland full of incompatible and proprietary boot systems and requiring custom kernels, device drivers, hardware trees, etc.
How would a closed proprietary system from Apple produce different results here?
GPUs are one of the main offenders. I messed around with linux based SBCs for a while and they end up being stuck with a particular kernel and distro most of the time. The binaries just won’t play nicely with anything new and update to date. It’s a real shame and I don’t see anything fixing it.
Right now the ARM landscape outside of this new Apple ecosystem
Congratulations! You have solved the standards problem [xkcd.com]! "this new Apple ecosystem" is just the n+1th version of how to boot ARM. And I wouldn't be surprised if Apple doesn't make it easy for other OSes to boot on their devices - we already have a good standard for booting on x86_64, UEFI... and then we have the Apple version of EFI. And if Apple doesn't care about other OSes, why would they even release sufficient hardware information to create reliable and high-performance drivers for other OSes?
If they can provide more performance for less money, that's great. I won't be buying one but I bet this will produce consumer-friendly pressure on MS, Intel, and AMD.
Do their margins go way up? If these ARM MacBook are successful, I'd definitely give an arm Linux laptop a chance too. I'm really waiting for a laptop / tablet convertible with a good desktop os interface like Ubuntu/Linux mint, then a good tablet interface like Android. (Not looking for a chrome book)
ARM cores are not new, and multicore ARM CPUs are not new. What is new is the unified memory archictecture, and moving that RAM from the motherboard into the CPU package.
In package memory may not be new (system-in-package chips are often used in low-cost, low power applications like MP3 players and cheap phones), but doing this for a large RAM sizes (8-16GB) on a high performance CPU is new. On of the most common system performance bottlenecks is memory performance. Making sure your machine has plenty of fast memory is often cheap way of improving system performance. Apple have effectively removed this bottleneck by moving the RAM in package. Apple have been doing package-on-package for several years, but moving the RAM is the next evolution. In a traditional motherboard, traces to RAM slots/chips may be tens of millimeters long (or longer). In the M1, it is tens of microns. This short distance allows the RAM to run much faster: 4.266GHz in the M1, as well as run at a much lower voltage.
Of course the downside is that the CPU is stuck with only 8GB or 16GB or RAM and you can never upgrade. That may not be a problem in the MacBook Air and Mac Book Pro, as those have not had upgradeable RAM in a long time, but Mac Mini users will be out of luck if 16GB isn't enough for their needs.
I do wonder what the next generation of CPUs will bring for the iMacPro and MacPro.... I cannot see them offering 32GB or more of in package RAM. Will they instead use the in-package RAM as a cache instead and have another memory controller for external RAM?
Why are you here? I wouldn't use Windows to save my life but I'm wasting my free time trolling Windows forums to tell people who aren't interested. These Apple stories attract people who wouldn't use Apple product to save their lives like a bug zapper attracts insects. There are an order of magnitude more people here hating on Apple than there are Apple 'hipsters'.
Motives aside, and Mac OS Big Sur issues notwithstanding, I am holding out replacing my 4 year old MBP for the 15 (16?) inch replacement with the new Apple Silicon. The things I care about are battery life and usability, both of which Apple smokes Windows at. A decent laptop starts at a grand if you expect to use it comfortably for 4 to 5 years, as I have done since the PowerBooks.
Note to people hating on Apple because the ecosystem just leapfrogged Wintel, these Apple stories are not aimed at you. No one
I agree, my 2017 hasn't lost keys but when i type on it my wife says, "what is that sound?" It sounds like there is something sticky under the keys even though i try to blow it out almost every time i use it. It's just uncleanable.
Frankly, that sounds like an abusive cult relationship. Not being sarcastic or anything. This really isn't healthy, and intentional or not (oh come on, they've got probably more psychological experts for this exact thing than other companies have employees!), it's one of those crimes that are extremely hard to prove. Not only because of many people's attitude of not believing such manipulative things can really happen, even though they happen all the time.
If I set here and stare at nothing long enough, people might think
I'm an engineer working on something.
-- S.R. McElroy
Meh. Lacks software library (Score:2, Interesting)
What apps do not run? (Score:5, Informative)
Meh. Who cares about speed when most of my apps won't run in MacOS
Thanks to the Rosetta 2 translation layer, they pretty much all run (there are a few minor glitches so far). What apps specifically do you use that you think will not run?
Re: (Score:2)
Logitech Utility
WinAmp
Waterfox Classic
And most of my Steam library, specifically Halo: MCC and Dark Souls 2
Re: What apps do not run? (Score:2)
Corsair, logetich, and Winamp all have Mac versions...
Re: (Score:2)
Winamp? Where? I don't see it on http://winamp.com/ [winamp.com].
Re: (Score:3, Insightful)
The only valid items on this list is your Steam library. Mac laptops were never suitable for gaming to begin with.
Re:What apps do not run? (Score:4)
Re: What apps do not run? (Score:2)
Re:What apps do not run? (Score:4, Interesting)
However, with all that said, macOS is currently the most secure desktop OS by default through isolating GUI apps (like what Wayland aims to achieve on Linux) and applying a sandbox to almost all apps which acts similarly to apparmor+seccomp-bpf but with a few creature comfort helper tools. It even has basic mandatory access controls to stop apps snooping in private user areas and will soon be the only system that has a truly robust Secure Boot implementation alongside a tamper-proof OS partition (a bit like implementing BSD Securelevels and making
Honestly, the security part of my brain has a wet dream over some of what Apple has seamlessly added to mainstream desktop computing but ultimately the dominant part of my brain which wants 'one computer to rule them all' will always opt for Windows. All I'm missing is proper telephony+SMS+4G baseband support and my desktop PC really can do everything with no need for any of this smartphone, tablet or laptop poop.
Re: (Score:2)
The first 3 aren't worth running to start with. Waterfox has a Mac version. I think Logitech and Winamp actually has Mac versions too.
The only valid items on this list is your Steam library. Mac laptops were never suitable for gaming to begin with.
At the same time, since these Macs will be able to support adapted iPad and iPhone games, it may get more gamer friendly. It still won't be "PC master race" when it comes to games, but I am sure most hardcore gamers probably have a custom speced computer anyhow.
Re: (Score:2)
The only valid items on this list is your Steam library. Mac laptops were never suitable for gaming to begin with.
And yet, some of them will probably run with WINE, and all will run with Boot Camp or in a virtualization environment. As a non-gamer, it's not a big deal to me, but for some people, it might be.
Re: (Score:2)
Mac laptops were never suitable for gaming to begin with.
I've gotten quite a bit of mileage out of my steam library on a 50gb bootcamp partition using a lowly 2018 macbook air that I picked up at launch. I use a desktop for the latest and greatest, of course, but my air serves me well for both work and leisure. I'll probably just move to a PC laptop and run linux with a windows partition next work laptop purchase.
Re: (Score:2)
Re: (Score:2)
I assumed he was talking about Windows libraries. That raises an interesting question; how well does WINE run in this environment?
Re: (Score:2)
Here's my understanding. Yes, WINE runs on ARM, but it does not do any kind of processor emulation. That means if you take your Windows source code and compile it to target ARM, your binary should work with WINE running on an ARM computer.
Now it's been years since I've used MacOS much, but I do remember the first version of Rosetta, which allowed the old PowerPC binaries to run on new Intel macs. When you opened an app for the first time it would transcompile the PowerPC object code to Intel object code
Re: (Score:3)
Two problems I read about... (Score:2)
I looked into the WINE thing a bit for ARM, from what I read there were two problems:
1) Metal didn't have enough features to allow for proper DirectX emulation.
2) For older games Rosetta 2 doesn't transcode 32 bit binaries.
Re: (Score:2)
Re: (Score:2)
There has been an android version of wine for several years.
I have not used this but it seems to support x86 and windows arm binaries.
Re: (Score:2)
Re: (Score:2)
All x86 VM apps...you know, the apps that make the Mac the best software development platform according to you.
Any Apple do not approve (Score:2)
What apps specifically do you use that you think will not run?
The question you should be asking after all the launch problems with Apple's certificate approval server failing, is what apps in the future will Apple let me run?
The thing that got me originally buying macs was that they ran intel so if I found that there was an app I needed I could boot it into Linux or even Windows. With the new M1 chip all you have is MacOS and nothing else. So if you need to use Linux or Windows for anything that's it, game over you have to buy a PC. The huge price increases, remov
Re: (Score:2)
Meh. Who cares about speed when most of my apps won't run in MacOS
So you are a Windows user?
Re: (Score:2)
Not necessarily... And there's WINE
A list of his apps that won't run on MacOS and that he can't find an alternative for would be more interesting than unsubstantiated claims. Especially in view of the fact that the vast majority of FOSS will compile and run on macOS. It used to be the rule around here that if you can't support your claims you were assumed to be full of shit until you did.
Not sure why expectations were even low... (Score:5, Interesting)
There have been comparisons for a while now, that showed modern day iPads were performing as well as, or better than a number of laptops...
So talking about a similar chip but with performance scaled up to a laptop with active cooling (in the case of the MacBook Pro), and to me it's not a surprise that people are seeing the performance Apple claimed they would.
Some of the people I follow on social have been showing equally impressive results on tasks across the board, weather it be hight end 4k video encoding, or compilation tasks for large projects.
For anyone that works day to day on a Mac it is VERY tempting to get even just the Mac mini for a performance boost, even with the 16GB limitation. Will probably wait a while before I take the plunge myself but it's great to see that we are at the dawn of a really promising line of systems that might actually deliver a significant increase of performance for real-world tasks, unlike the last few years of intel updates where often it was indiscernible if anything was really faster at all.
Re: (Score:2, Informative)
Even in Geekbench, which seems to be tuned to favour Apple parts, the M1 is lagging behind Intel Macbooks. And Macbooks are not fast compared to laptops with better cooling, and certainly not compared to AMD laptops.
So while it's surprisingly decent the fact is it takes longer to export a video in Final Cut on an M1 than it does on an older Intel Macbook Pro. The GPU is similar to mid range AMD and Nvidia mobile chips from 3-4 years ago.
Re: (Score:2)
Are we looking at the same Geekbench scores? I see the M1 using x86 emulation beating the native Intel Macbooks, and pulling out far ahead using the native Geekbench.
Re: (Score:2, Informative)
Still pushing that bullshit, jcr? Rosetta is an emulator, it cannot be anything other than an emulator.
Translation and emulation are orthagonal, not mutually exclusive. You've been told that countless times, but you continue to repeat the same garbage because you're incapable of learning anything.
Interpreter, jcr, what Rosetta isn't is an interpreter. It is an emulator because it emulates an x86 Mac. Anyone technically competent would understand this, but not you.
Re: (Score:2)
But Wine is not an Wine is not an Wine is not an Wine is not an Emulator?
Quack!
Re: (Score:3, Insightful)
Right, and in Cinebench we see the Rosetta scores generally matching the previously used Intel chips, and the native scores significantly surpassing them. Yes, they're not as good as Zen 3, but that's not relevant since Apple has never shipped a Zen 3 CPU. The two questions here are, are the M1 macs an improvement over their predecessors, and how much of a performance penalty does the lack of native apps have. And on the former, the answer is "a rather large improvement", and on the latter, the answer is "2
Try reading the links (Score:4, Informative)
Even in Geekbench, which seems to be tuned to favour Apple parts, the M1 is lagging behind Intel Macbooks.
Not from any of the reports I've read.
Just one example is TheNextWeb article [thenextweb.com] which states:
The Apple Silicon Geekbench results also blow the 2019 16" i9 MacBook Pro out of the water. That scored: 1118 / 6762.
The Apple Silicon Geekbench results for the new systems (which remember are much lower specced normally than a 16" MacBook Pro):
Mac Mini M1: 1682 / 7097
MacBook Air M1: 1687 / 7433
MacBook Pro (13") M1: 1714 / 6802
So every single new M1 Apple system is beating one of the top of the line Apple Intel systems...
Re: (Score:2)
Why is the Pro lagging behind the Air?
Re: (Score:3)
Why is the Pro lagging behind the Air?
It did on the very first released benchmark results, and never again after that.
A possible explanation: Macs do a lot of work when they are started for the first time (like collecting information so that "search" works instantly), so if you start a brand new Mac, download a benchmark, launch the benchmark _immediately_, it will run slower because your Mac does other things.
With all newer M1 Macs, the order is MacMini consistently very slightly faster than MacBook Pro, and MacBook Air equal to MacBook
Re: (Score:2)
What does Intel have to do with the M1? Why ask them?
Re: (Score:3)
From the original poster:
Mac Mini M1: 1682 / 7097
MacBook Air M1: 1687 / 7433
MacBook Pro (13") M1: 1714 / 6802
So, the MacBook Pro is slower in multicore than the Air (6802 vs 7433), hence the question. And both are marked "M1".
Re: (Score:3)
The A series chips after 3-4 minutes just tanked in performance to about 60%, and continued to drop until about 50%ish performance. While this might be fine on a consumption device like a iPad, they aren't on a laptop that is expected to work for more then a few moments at a time.
Its also one of those things that I'm noticing is lacking in these reviews. What
Re:Not sure why expectations were even low... (Score:5, Informative)
Re: (Score:2)
How long does it take to run? Does CPU 2017 take longer than CPU 2006? Run time depends on the system, suite, compiler, tuning, and how many copies or threads are chosen.
https://www.spec.org/cpu2017/Docs/overview.html#Q11
So according to the link you gave me, it will depend on how they ran the test. It says that if you run each test alone, they can be done in a few minutes.
Can you please show/quote where anandtech states how they ran the benchmarking? Was it a full suite that will take hours, or a single test that is done in a few minutes and then lets the system cool down? I tried looking myself, but can't find any indication how the tests were run
Re: (Score:3)
So according to the link you gave me, it will depend on how they ran the test. It says that if you run each test alone, they can be done in a few minutes.
Can you please show/quote where anandtech states how they ran the benchmarking? Was it a full suite that will take hours, or a single test that is done in a few minutes and then lets the system cool down? I tried looking myself, but can't find any indication how the tests were run
If you believe you have found such a fundamental flaw in Anandtech's testing, then you might want to inform them of your concerns. I'm sure they would be pleased to have their errors pointed out so they can correct them.
Re: (Score:3)
But once again those are artificial benchmarks. Where are the benchmarks of actual applications that people use? The only one in your linked article is Tomb Raider, all the rest are artificial. They didn't even try Cinebench.
Rise of the Tomb Raider, enthusiast
M1 - 39 FPS
Low end Ryzen 4600H with 1650 - 58 FPS
High end i7 10th gen with an old 1660Ti - 82 FPS
Unfortunately they don't give 95 and 99th percentile scores.
Re: (Score:2)
For anyone that works day to day on a Mac it is VERY tempting to get even just the Mac mini for a performance boost, even with the 16GB limitation. Will probably wait a while before I take the plunge myself but it's great to see that we are at the dawn of a really promising line of systems that might actually deliver a significant increase of performance for real-world tasks, unlike the last few years of intel updates where often it was indiscernible if anything was really faster at all
While I am tempted by the speed and the novelty of the new, the one thing missing for me is being able to run intel based VMs. And I believe that Apple has said that Rosetta 2 doesn't do that.
I'm almost to the point of replacing my 2014 mac mini with an intel 2018 model, rather than new shiny. And I'm sure a lot of people are in the same boat.
Re: (Score:2)
As long as those "real-world tasks" don't require much memory or scaling to multiple cores or x86 OS compatibility, otherwise known as "real world tasks that suit a low power portable processor". But sure, cherry pick the results, as you always do SuperKendall.
Re: (Score:2)
Re: (Score:2)
or compilation tasks for large projects.
Can this even compile large projects? AFAIK chromium for example requires more than 16GB of ram to compile.
Apple announces Apple fans love (Score:3, Funny)
Re: (Score:2)
Yeah, I've noticed that most of the reviews that have come out so far have been Apple fanboys who have been bragging how well it runs Geekbench.
I don't know about you, but I don't use Geekbench on a daily basis. I do use Docker, though... does that even work in a Mac with an M1 processor?
Re: (Score:3)
I don't know about you, but I don't use Geekbench on a daily basis. I do use Docker, though... does that even work in a Mac with an M1 processor?
not yet, it Docker says it's coming [appleinsider.com]
Re: Apple announces Apple fans love (Score:2)
But it's the ARM version of Docker (running ARM containers images & ARM app binaries), so probably not helpful unless you have ARM versions of your programs or ability to build them from source.
Not part of the cult... (Score:2)
Sorry, none of the apps I need exist in the Apple ecosystem.
The performance doesn't impress me.
They were on the right track with the Intel Macs- and made a mistake by eschewing AMD's new chips for ARM. Big mistake.
Re: (Score:2)
Sorry, none of the apps I need exist in the Apple ecosystem.
Just out of curiosity, what apps are those?
Re: (Score:2)
Android-only ones maybe?
I doubt it. For one thing this is a Mac processor not a iPhone/iPad processor so mobile apps aren't really an issue in this context:
Secondly, most of the apps worth using are already available on Android and iOS since iOS users are more profitable than Android users. Finally pretty much all FOSS is compilable on MacOS and iOS and even most of the Microsoft products are available on MacOS these days. I've been using L
Cost based analysis also is awesome for Apple (Score:5, Informative)
Re:Cost based analysis also is awesome for Apple (Score:4, Insightful)
How about a cost benefit analysis for the buyer?
Non upgradable RAM, mid range performance for high end prices...
Re:Cost based analysis also is awesome for Apple (Score:5, Insightful)
How about a cost benefit analysis for the buyer?
Non upgradable RAM, mid range performance for high end prices...
Apple, as we all know, does "value" based pricing, not cost-up. And value = brand, apps, ease of use, etc. They are not optimizing for folks who want upgradable RAM, etc. From apple's perspective, those folks can buy the high-end MacBooks ... And $40B in revenue per year suggests their product strategy is working.
Re:Cost based analysis also is awesome for Apple (Score:5, Insightful)
You are 100% right about value-based pricing, but slashdot readers have never understood and will never understand that. This is a crowd that thinks computers should be priced be weighing the silicon, metal, and plastic that goes into their production and charging a flat rate plus 5% markup.
Re:Cost based analysis also is awesome for Apple (Score:5, Interesting)
The money they don't have to contribute to Intel's net profit isn't even the main reason for the transition. By going to their own CPUs, Apple's product release cycles aren't subject to a third party's schedule.
-jcr
Re: (Score:2)
A paradigm shift happening in front of us (Score:5, Insightful)
If you hate Apple enough, it's easy to dismiss what's happening right in front of us at the moment. What I find most interesting is looking at what this M chip can do already but also to keep in mind that they have yet to release their "Pro" spec chip (P-series?).
This is just the first iteration of a mobile level chip and it's already outperforming x86 CPUs and discrete GPUs, albeit older ones. Going the route of previously unseen levels vertical integration also opens up all sorts of new systems development avenues in areas that are up until now unexplored by PC makers.
Given the performance levels we're already seeing, I believe we're all in for quite the shock when they release their P-series chip.
Re: (Score:2)
They are going to have to work hard to overtake AMD mobile chips. Take this Cinebench test for example: https://m.hexus.net/tech/news/... [hexus.net]
The M1 scores less than half what the AMD mobile part does. The GPU is years behind current AMD ones as well.
Re: (Score:2)
Update:
It looks like the Bits and Chips CInebench R23 result is not for the Apple M1 SoC but from an A12Z. MacRumours has a result in that it reckons is from a genuine new M1 13-inch MacBook Pro, though, and it is somewhat better than the A12Z, of course. It quotes Twitter user @mnloona48_ who shared an MBP M1 unboxing and various other tests on his new 8GB machine. Mnloona48's Cinebench R23 scores are an impressive 1,498 single core, 7,508 multi core.
Re: (Score:2)
Which if you look at the chart is less than 3/4 the mid range AMD mobile CPU.
Re: (Score:2)
Anandtech says that the M1 scored 1522 (single) and 7833 (multi), using Cinebench R23.
Mac Miini M1 tested [anandtech.com]
There are also specbench 2006 results, though some of the more spectacular results are due in part to the optimizing compiler.
Re: (Score:2)
Now I'm impressed. My year old Linux box with a Ryzen 3400G is 1068 single and 4810 multi. And it is entirely fast enough for my use.
The remaining problem with the mini is the lack of ports. I have four usb things plugged in now, and I would need two more to make up for the second drive and the DVD drive. And I still need a plug for a flash drive, so I'd need 8 USB ports on the mini.
Can't argue the processor performance though.
Re: A paradigm shift happening in front of us (Score:2)
That article you linked to has an update with M1 numbers a lot closer to the AMD part.
Re: (Score:2)
You are right, the updated score is a little under 3/4 the CPU performance of the mid range AMD part.
Read the Links Luke (Score:4, Informative)
The M1 scores less than half what the AMD mobile part does.
Not if you read the Ars link from the summary [arstechnica.com]
Scroll down to the Cinebench R23 section...
The M1 is out-performing the Ryzen 7...
Re: (Score:2)
Only against the low-mid range ultra low power 4700U. It's only got a 15W TDP, very much a budget low power device.
The link I posted compared with a mid range 4800 with 35W TDP.
By the way laptops with a 4800 start at around £500 and have a much better GPU than the M1.
Re-Read Your Own Link, Luke (Score:2)
The link I posted compared with a mid range 4800 with 35W TDP.
The article at the link you posted has been updated [hexus.net], the M1 is better at single thread and almost the same in multi-threaded performance... on Apple's lowest end laptop. Compared to both mobile and desktop CPU's from AMD.
Re: (Score:2)
Unless less points are better on this benchmark, or you think 10,600 is less than 4,530 ... Your link shows the Ryzen 4800 is a fair bit ahead of the M1. Although I looked up the same benchmark (Cinemark R23) on a different site and it showed them nearly tied.
No idea what's going on there, but my Ryzen 48000H system cost me $1100 and it came with a discrete, gameable GPU (RX 5600M) whereas the $1000 Apple does not. I've had it for 6 months, the price has already dropped since then. This month AMD released t
Re: (Score:2)
Re: (Score:2)
It is actually quite easy to predict. Expect single core performance to be no more then 5% faster then the M1. Better cooling will facilitate a slightly faster clock rate but the core should be the same. For multi-threaded performance expect ~1.75 times faster with 8 fast cores or ~3.3 times faster for 16 fast cores.
The unknown is the memory architecture, GPU, and PCIe lanes. The internal GPU might be excluded in favor of more CPU cores. The number of memory lanes is unknown but it is possible we se
Re: (Score:3)
Agreed. I think the P-series is going to be a game changer for the desktop folks who can really use it.
Right now my concern is the complete lack of virtualization on the platform. The best you can do is an arm-based linux distro. Both Parallels and VMware have reported that they are working on a version of their platforms that will work with the Apple silicon, but I can't imagine that this isn't going to suck as there would have to be a lot of instruction set translation going on.
I have no idea how this
Remember the PowerPC Macs? (Score:2)
This is just the first iteration of a mobile level chip and it's already outperforming x86 CPUs and discrete GPUs, albeit older ones.
You have a short memory. Apple used to use PowerPC before switching to Intel. They were RISC-based and also claimed to be much more powerful than Intel CPUs...and yet Macs had a terrible market share. There was a reason that Macs took off when they switched to Intel.
Re: (Score:2)
Not saying its not good, just not buying the first (Score:3)
As in most of, well anyting that is complex and new technology, I'll allow others to buy the first ones and work out the kinks.
Simple rule for me: Never buy the first generation of technology. I've worked the back side of technology and see how they test things, or just don't...
Re: (Score:2)
I think, for a lot of people, there's just not a compelling case to be on the leading edge with new tech anymore. We've been at a point for a number of years where the performance of even relatively "old" tech outstrips what most people need.
I can't say that I ever feel constrained writing code on this 2015 MacBook Pro, for instance.
Palm Foleo, Samsung ARM Chromebook, MacBook M1 (Score:2)
I am glad to finally see a competitive ARM laptop.
Though one day we may see the Apple MV chip, based on Risc-V, to avoid ARM royalties.
Is this an ad? (Score:3, Insightful)
All the comments compare the M1 to the writer's own "expectations", which is meaningless. Actual objective benchmarks show it's nowhere near a decent gaming / workstation PC. Couple that with Apple's walled garden, anti-repair and anti-recycling stance, and terrible ethics, and it's a very solid "no".
Re: (Score:2)
Hey, don't insult ads like that!
Hope this will spur the ARM world standardize (Score:5, Interesting)
I hope that Apple's success with ARM will spur the rest of the ARM industry to get its act together and get ARM standardized into a complete and compatible platform that Linux and Windows can work with. Right now the ARM landscape outside of this new Apple ecosystem is quite a wasteland full of incompatible and proprietary boot systems and requiring custom kernels, device drivers, hardware trees, etc.
When we get to the point that a person can buy an ARM laptop, server, or desktop mini PC, and then go to RedHat.com and get stock RHEL for it, or go to Fedora, Ubuntu, or Arch, and get a stock distribution for it, or even a Windows 10 ISO from Microsoft, then Intel start to find themselves in trouble. Right now they have little to worry about in the PC world, other than AMD. Except for some hobby or embedded applications, the way things are currently, I would pick an Intel or AMD CPU any day over ARM. And that includes my router firewall. I want that to change!
It's not that ARM can't perform. It's that each ARM system is a world of its own right now. I've heard about some standards being proposed and attempted, but I've not seen much evidence of movement there. Currently most vendors are satisfied to supply binary blobs for proprietary Android installations.
Re:Hope this will spur the ARM world standardize (Score:5, Interesting)
Right now the ARM landscape outside of this new Apple ecosystem is quite a wasteland full of incompatible and proprietary boot systems and requiring custom kernels, device drivers, hardware trees, etc.
How would a closed proprietary system from Apple produce different results here?
Re: (Score:2)
It's a much bigger player. Might spur the rest to get together. I can dream.
Re: (Score:2)
GPUs are one of the main offenders. I messed around with linux based SBCs for a while and they end up being stuck with a particular kernel and distro most of the time. The binaries just won’t play nicely with anything new and update to date. It’s a real shame and I don’t see anything fixing it.
Re: (Score:2)
Right now the ARM landscape outside of this new Apple ecosystem
Congratulations! You have solved the standards problem [xkcd.com]! "this new Apple ecosystem" is just the n+1th version of how to boot ARM. And I wouldn't be surprised if Apple doesn't make it easy for other OSes to boot on their devices - we already have a good standard for booting on x86_64, UEFI... and then we have the Apple version of EFI. And if Apple doesn't care about other OSes, why would they even release sufficient hardware information to create reliable and high-performance drivers for other OSes?
Cool (Score:2)
Re: (Score:2)
You forgot the sarcasm tag there. :)
Or did you mistype "consumer-unfriendly".
Because if they want to imitate Apple... and they usually do... sadly...
no trust = no sale (Score:2)
What's apples profit margins on new architecture? (Score:2)
mac mini (Score:5, Funny)
Re: (Score:2)
Re: (Score:3)
Commenting on Apple's price-gouging markups in 2020 is rather tired. Didn't they start this in the mid 00s?
the virge? (Score:4, Funny)
you mean the morons that do not know the difference between a zip tie and tweezers? Yea Ill trust their "expert" opinion
In-package RAM is the key (Score:5, Informative)
ARM cores are not new, and multicore ARM CPUs are not new. What is new is the unified memory archictecture, and moving that RAM from the motherboard into the CPU package.
In package memory may not be new (system-in-package chips are often used in low-cost, low power applications like MP3 players and cheap phones), but doing this for a large RAM sizes (8-16GB) on a high performance CPU is new. On of the most common system performance bottlenecks is memory performance. Making sure your machine has plenty of fast memory is often cheap way of improving system performance. Apple have effectively removed this bottleneck by moving the RAM in package. Apple have been doing package-on-package for several years, but moving the RAM is the next evolution. In a traditional motherboard, traces to RAM slots/chips may be tens of millimeters long (or longer). In the M1, it is tens of microns. This short distance allows the RAM to run much faster: 4.266GHz in the M1, as well as run at a much lower voltage.
Of course the downside is that the CPU is stuck with only 8GB or 16GB or RAM and you can never upgrade. That may not be a problem in the MacBook Air and Mac Book Pro, as those have not had upgradeable RAM in a long time, but Mac Mini users will be out of luck if 16GB isn't enough for their needs.
I do wonder what the next generation of CPUs will bring for the iMacPro and MacPro.... I cannot see them offering 32GB or more of in package RAM. Will they instead use the in-package RAM as a cache instead and have another memory controller for external RAM?
Re: (Score:3, Insightful)
M1 + Apple Baggage = No Thanks
Why are you here? I wouldn't use Windows to save my life but I'm wasting my free time trolling Windows forums to tell people who aren't interested. These Apple stories attract people who wouldn't use Apple product to save their lives like a bug zapper attracts insects. There are an order of magnitude more people here hating on Apple than there are Apple 'hipsters'.
Re:Not Yet (Score:4, Funny)
Yep, the flaming fanbois (both parties) are out tonight, just making shit up to support their party. What a waste of story.
BTW, I hear the M1 voted for Trump...
Re: Not Yet (Score:3, Insightful)
Note to people hating on Apple because the ecosystem just leapfrogged Wintel, these Apple stories are not aimed at you. No one
I've found the opposite to be true (Score:2)
Everything is more expensive in the Apple ecosystem, development costs, maintenance costs, deployment costs
I've not found that at all, because Apple stuff simply lasts a lot longer than any Windows gear I have used.
The fact that my MacBook Pro 15" from 2013 is still usable for day to day development says it all.
How are maintenance costs higher if the systems last much longer?
your choices are very restricted.
Looking at Apple's lineup, I really don't see where any such restriction exists.
Also, with more peopl
Re: (Score:3, Funny)
The fact that my MacBook Pro 15" from 2013 is still usable for day to day development says it all.
Wish the same were true for the ones from 2017.
[Insert photo of a computer with five missing butterfly keys here]
:-D
Re: I've found the opposite to be true (Score:2)
Re: (Score:2)
Frankly, that sounds like an abusive cult relationship. Not being sarcastic or anything. This really isn't healthy, and intentional or not (oh come on, they've got probably more psychological experts for this exact thing than other companies have employees!), it's one of those crimes that are extremely hard to prove. Not only because of many people's attitude of not believing such manipulative things can really happen, even though they happen all the time.