Which engine do you use that runs under macOS/M1 as well as under PC? With Rosetta 2?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Developer Delves Into Reasons Why Apple's M1 Chip is So Fast
- Thread starter MacRumors
- Start date
- Sort by reaction score
Which engine do you use that runs under macOS/M1 as well as under PC? With Rosetta 2?
I was just reading through the comments because that is exactly how it appears to me.
Not that it matters how they got there - in many ways - just where they are.
To add to the AMD situation , AS is on 5nm , AMD have no such products , so they are not competing over capacity , 2nd AMD preferred the Xbox/PS orders over the DIY crowd , which is the 2nd bottleneck in supply.
Sure, my first Mac II had separate chips for the integer CPU 68020 and the floating-point FPU 68881, and didn't even ship with a memory management unit they included in later successors. All that is now just part of a single core packed onto single piece of silicon that includes a bunch of cores, an entire graphics card, neural engines, i/o controllers, and who knows what else that used to be sprawling dozens of chips across a foot-wide PCB. It's still smaller, faster, cheaper, and more efficient than all that mess used to be.
I sold my mid 2015 15" MBP for $790 through Mac Me An Offer. I almost got screwed by trading it to Apple for what you mentioned above. Thankfully someone on this forum told me about Mac Me An Offer and that's how I got $790. I'd highly suggest you do the same.
No one is arguing that Intel is better, I think all of us who aren't impulse-buying this thing are simply arguing "it can't run the (Windows) software that I need".
Let's see how things progress. Hopefully in a year or two there will be serious development with Windows on ARM that allows it to run on these Macs, or M2 Macs, as well as supporting x86 programs through emulation.
Looks like Intel and AMD fanboys are butthurt that Apple M1 is destroying the competition.
All I heard is Intel/AMD has faster chips ... but whataboutism R7 4800U, only 7nm, wait until xyz comes out with better stuff!
Why can't fanboys just admit having an M1 chip at 10W TDP (including 8core-CPU, 8CU-GPU, Accelerators, DRAM, TB/USB controllers, SSD controller) performing near desktop speeds, it's kinda radical?
Don't forget video reviews of Apple M1 with 65W CPUs ... just to show how "inadequate" apple is at replaing PCs . LOL
All of a sudden, fanboys who have been poking fun at Apple for years about their pricing, heat issues, and "lack of innovation" just want the facts to care about their feelings.
Goes to show you that most people have high levels of confirmation bias.
Haters are gonna hate. They will always find ways to complain and bitch (right to repair, price, larger battery, macos, blah blah blah ...)
All I heard is Intel/AMD has faster chips ... but whataboutism R7 4800U, only 7nm, wait until xyz comes out with better stuff!
Why can't fanboys just admit having an M1 chip at 10W TDP (including 8core-CPU, 8CU-GPU, Accelerators, DRAM, TB/USB controllers, SSD controller) performing near desktop speeds, it's kinda radical?
Don't forget video reviews of Apple M1 with 65W CPUs ... just to show how "inadequate" apple is at replaing PCs . LOL
All of a sudden, fanboys who have been poking fun at Apple for years about their pricing, heat issues, and "lack of innovation" just want the facts to care about their feelings.
Goes to show you that most people have high levels of confirmation bias.
Haters are gonna hate. They will always find ways to complain and bitch (right to repair, price, larger battery, macos, blah blah blah ...)
Is anyone able explain why M1 struggle hard in one instruction cpu simulation in Gary Speed test G who flies in the rest?
It took about 22 seconds to M1 while 7 to Intel.
test itself starts about 4:45
It took about 22 seconds to M1 while 7 to Intel.
Last edited:
Ill give it a go (2 reasons), when a SoC is in idle (which is the case in this kind of single instruction test) it clock/power gates domains , the more aggressive you power down when doing nothing if its by how long you wait before starting to shutdown or by how many domains you power gate (or both) the worst you will perform in such a useless test , my bet is that M1 does way more powering down cycles and probably to deeper states as well in this test then the intel counter part , so it pays more performance penalty for the overall "test" , note this is one of the most useless tests ever conceived as it will penalize a good power efficient design while rewarding an "always on" CPU design , where in real life there is no such use case.
Another reasonable guess , is that the SoC will not go to a high performance mode when doing single instructions , as the expectation is that you can go to sleep ASAP , which is reasonable to assume when you have a single instruction in your pipeline , again another reason why this test is as useless as it gets!
As the reviewer didn't show us the frequency in which the test ran at this makes the comparison futile , hope it helps!!!
just to ease your mind , there is no way Intel does any single instruction 3x faster then Apple (maybe aside from the AVX instructions , that needs to be rewritten as several NEON instructions , but AVX is barely used anyway)
That’s not been true since the 1980’s. (And I made my first money from programming on a CDC Cyber 175. _That_ was a reduced instruction set computer with just 60 instructions. Not regular because instructions were 15, 30 or 60 bits).
Isn't Tiger Lake the answer to the M1? It is in the Dell XPS 13" right now.
https://www.forbes.com/sites/brooke...nchmarks-13-inch-macbook-pro/?sh=66aa32c967e1
Last edited:
There is more to a computer than how fast the CPU is. The M1's that are out now are the entry level CPUs.
Do you have a link?
I question Apple's influence because I had Amigas between 1988 and 1994 and it seemed pretty clear to me that good as the Amiga was, it was hampered by Commodore's financial problems and hence they never got beyond the 68020 in terms of a domestic popular machine - the bigger Amigas dont really count as they were not mass market.
I remember David Braben (Elite/Frontier) remarking how much more powerful the 486 DX chips were and how there was little point continuing to develop for the Amiga when the PC just crushed it from a power perspective.
There was never any references to Apple who at that time had firmly gone down the DTP route as a means of clinging on to life.
you do realise that you come across as the very worst kind of zealot yourself with posts like that.
I'm sure it does, but I just wouldn't lose any sleep myself worrying about a resale value. I buy a computer to use now and for however long I need it for. I'd rather not deal with people who want to buy one second hand from me - too many nutters.
It actually is a system on chip. The term is used to refer to a design methodology where blocks are independently designed and communicate on a bus rather than via specialized connections. And each of the circuits mentioned (cpu, gpu, neural engines, secure store, memory controller, etc) are indeed on the same silicon chip. The only thing that’s on a separate chip in the package is the RAM.
SOC is used to differentiate from “ASIC,” which is a methodology where each of those blocks would have its own unique interface, and changes made to one block would have effects that affect the others.
M1 vs Zen 4 will be the more interesting comparison.
Sure Zen3 vs M1 is the more practical comparison, because they are released at the same time.
However, they use manufacturing nodes a generation apart. Apple, with tiny SoC's and sky-high margins can afford to be the leading 5nm customer, while amd had to settle for 7nm. A fair bit of that Perf-Per-Watt superiority of the M1 comes down to that more advanced manufacturing node.
How much? That's the interesting part and remains to be seen. Especially on a 45W+ design.
And no Zen4 will NOT beat the M1's perf-per-watt ratio.
Sure Zen3 vs M1 is the more practical comparison, because they are released at the same time.
However, they use manufacturing nodes a generation apart. Apple, with tiny SoC's and sky-high margins can afford to be the leading 5nm customer, while amd had to settle for 7nm. A fair bit of that Perf-Per-Watt superiority of the M1 comes down to that more advanced manufacturing node.
How much? That's the interesting part and remains to be seen. Especially on a 45W+ design.
And no Zen4 will NOT beat the M1's perf-per-watt ratio.
Weird you talk about Apple being a closed system and then mention Ps5 AMD XseX are cheaper... those 2 are even more close than Apple. And the reason the ‘machine’ is cheap is because they are heavily subsidised and you will eventually have to pay more because games are more expensive. So saying Ps5/XseX are cheaper are really for the short sighted people
given that a 7nm amd needs 66% more clock speed to match apple, apple should still hold the lead substantially even against zen3.
Nope. The Problem has always been technology: Manufacturing.
For decades there was a constant :
Intel's Manufacturing Prowess was the undisputed king. At least 2 generations ahead of everyone else.
Right now, intel's best is two generations behind TSMC's 5nm.
Yup, a majority of PC buyers don’t even know how much RAM they have or need, never mind even thinking about upgrading it. While there is definitely a good % of enthusanists would will upgrade parts every couple years, most people just buy a computer. Then in 4-5 years when it get ‘slow’, they buy a new one.