Apples new M1 microprocessor

[brucehoult]

Very true and exactly why I'm not buying one.

OCSP validation of binaries meaning rights to run software you purchased can be arbitrarily revoked, Linux probably will never work on the hardware apparently due to signed boot and no documentation so it's landfill immediately when macOS is dumped, half the dev tools out there aren't even compiled for ARM, docker doesn't work

As things are now, things will forever be.

[brucehoult]

Many Mac users work on technical fields, especially on ISPs, and they rely heavily on virtualization of Intel based Linux and FreeBSD systems. Those users also need a lot of memory, more than 16 GB.

There will never ever be a new Apple Silicon chip with more cores, more RAM.

[brucehoult]

Linux probably will never work on the hardware apparently due to signed boot and no documentation so it's landfill immediately when macOS is dumped,

Rather early to say that, many versions of Apple hardware have been poorly supported or even unsupported  by Linux when they first came out, and became perfectly tractable a few weeks/months down the line.

Exactly.

Production machines just started landing in user's hands this week.

There are a lot of very clever people who would love to see Linux, docker and other things running on these. Gentlemen, start your engines.

[brucehoult]

VMWare have had a preview of Fusion for ARM on the Mac available for a while, they are also running one of their public previews of ESXi for ARM based servers. So it cuts the other way too, an ARM based Mac may become the only option for developers who want to play with virtualised ARM on the desktop to support playing with ARM server farms.

VMWare dude has ESXi running on a Pi 4, wants to get it working on ARM Macs...

https://twitter.com/BruceHoult/status/1327540805947461632

[tooki]

[...]
Many Mac users work on technical fields, especially on ISPs, and they rely heavily on virtualization of Intel based Linux and FreeBSD systems. Those users also need a lot of memory, more than 16 GB.

So these machines won't work for them.

[...]

So, can we think of these things as decently performing but closed appliances, suitable for not-too-challenging applications by "aspirational" users that are not too price sensitive?   

No, they're quite high-performing (for most tasks) semi-closed appliances suitable for lots and lots of challenging applications for both consumers and professionals who are willing to spend a little bit more up-front on a computer in exchange for reduced hassle during its lifespan.

[SilverSolder]

[...]
Many Mac users work on technical fields, especially on ISPs, and they rely heavily on virtualization of Intel based Linux and FreeBSD systems. Those users also need a lot of memory, more than 16 GB.

So these machines won't work for them.

[...]

So, can we think of these things as decently performing but closed appliances, suitable for not-too-challenging applications by "aspirational" users that are not too price sensitive?   

No, they're quite high-performing (for most tasks) semi-closed appliances suitable for lots and lots of challenging applications for both consumers and professionals who are willing to spend a little bit more up-front on a computer in exchange for reduced hassle during its lifespan.

I thought this thread (including the Anandtech link earlier) has established that the new Apple SoC is primarily efficient, rather than being the ultimate in absolute performance - i.e. the Intel and AMD processors it is being compared with are able to draw much more power if needed, at which point they blow the new Apple chip away? 

I'm glad there are choices available from Apple, Microsoft, and the open source community.   As soon as you give up on the idea that one PC has to do everything, life becomes much simpler (but more cluttered)! 

[Marco]

I'm glad there are choices available from Apple, Microsoft, and the open source community.

For now.

Apple sucks up massive amounts of margins from the market though and now PCs won't be able to get at least some of that back through the CPU/GPU vendors. Also Microsoft has been floundering for a long time and even if they weren't the ever impossible task for Microsoft to truly provide a stable platform over millions of different PC configurations is a constant drag on their competitiveness relative to Apple.

On top of that Apple is getting in on finance in a big way and that will make using a Mac together with your iPhone even more of a package deal.

The M1 isn't more choice, it's a big part of the impending doom for "open" PCs. Microsoft needs to make huge changes to regain it's competitiveness in the consumer space and it needs to do it soon ... processor performance is the least of their problems (as long as they stop wasting time with ARM, Qualcomm ain't Apple).

[brucehoult]

I thought this thread (including the Anandtech link earlier) has established that the new Apple SoC is primarily efficient, rather than being the ultimate in absolute performance - i.e. the Intel and AMD processors it is being compared with are able to draw much more power if needed, at which point they blow the new Apple chip away? 

Nope.

The M1 is right up there, core for core, against the biggest baddest hot running desktop 125W i9 10900K, 105W AMD Ryzen 9 5950X etc.

It's got fewer cores, of course.

[bd139]

I'm glad there are choices available from Apple, Microsoft, and the open source community.

For now.

Apple sucks up massive amounts of margins from the market though and now PCs won't be able to get at least some of that back through the CPU/GPU vendors. Also Microsoft has been floundering for a long time and even if they weren't the ever impossible task for Microsoft to truly provide a stable platform over millions of different PC configurations is a constant drag on their competitiveness relative to Apple.

On top of that Apple is getting in on finance in a big way and that will make using a Mac together with your iPhone even more of a package deal.

The M1 isn't more choice, it's a big part of the impending doom for "open" PCs. Microsoft needs to make huge changes to regain it's competitiveness in the consumer space and it needs to do it soon ... processor performance is the least of their problems (as long as they stop wasting time with ARM, Qualcomm ain't Apple).

I don’t see this changing any time soon. There’s very little penetration of macOS outside the US and almost no corporate users in Europe. Anywhere with a strong security posture has windows as well due to GP and management infrastructure. The open PC runs the world.

The open PC probably also runs Altium or Zukon which design the Macs 

[magic]

Up against 5GHz turbo, tens of megs of L3 and AVX2?

 

[bd139]

Also against “affordable performance”

[borjam]

I don’t see this changing any time soon. There’s very little penetration of macOS outside the US and almost no corporate users in Europe. Anywhere with a strong security posture has windows as well due to GP and management infrastructure. The open PC runs the world.

The open PC probably also runs Altium or Zukon which design the Macs 

The funny thing about "open" is that it may have so many meanings!

I remember when Intel, while embroiled in antitrust troubles, boasted that their processors were "open" while others such as SPARC were "proprietary".

Notably, SPARC was an architecture specification and anyone was able to purchase a license, design SPARC compliant processors and have them certified. The very definition of proprietary!

[bd139]

Indeed.

PC platform and ISA is very open and well standardised. Intel are just sharecroppers on the land with everyone else at this point. They don't own it. The spiritual owner is IBM in theory.

Not that any of this matters as the only thing people care about is the software and that is portable in theory.

On that note, this is interesting: https://www.opengroup.org/openbrand/certificates/1212p.pdf

macOS on ARM is officially "a Unix".

[borjam]

On that note, this is interesting: https://www.opengroup.org/openbrand/certificates/1212p.pdf

macOS on ARM is officially "a Unix".

Yep, macOS has been Unix compliant for a long time.

So, is an OS with an open source kernel (and other open source bits) open or not?

Now imagine Apple adopting Motif!! That would be funny.

[bd139]

Exactly my point

At least Motif is consistent over time 

(Nightmares of Cadence coming back now)

[borjam]

Exactly my point

At least Motif is consistent over time 

You're not even wrong.

Open Look!

(I remember I used Open Look on FreeBSD for a while)

[bd139]

Ugh I never got on with that one. I had a SPARCstation 5 as a workstation with it for about 6 months. Annoyingly my home directory was network mounted and they had a script that rewrote everyone's X sessions overnight. I had a locally compiled FVWM available fortunately.

[SilverSolder]

I thought this thread (including the Anandtech link earlier) has established that the new Apple SoC is primarily efficient, rather than being the ultimate in absolute performance - i.e. the Intel and AMD processors it is being compared with are able to draw much more power if needed, at which point they blow the new Apple chip away? 

Nope.

The M1 is right up there, core for core, against the biggest baddest hot running desktop 125W i9 10900K, 105W AMD Ryzen 9 5950X etc.

It's got fewer cores, of course.

So in the same sense that a 4 cylinder engine is just as powerful as a V-8, cylinder per cylinder?

[DiTBho]

why Apple didn't choose RISV-V?

[bd139]

Because it doesn't really exist yet in the real world. And they have a long history of ARM usage and experience.

This is interesting: https://www.avantek.co.uk/arm-server-h270-t70/

[brucehoult]

I thought this thread (including the Anandtech link earlier) has established that the new Apple SoC is primarily efficient, rather than being the ultimate in absolute performance - i.e. the Intel and AMD processors it is being compared with are able to draw much more power if needed, at which point they blow the new Apple chip away? 

Nope.

The M1 is right up there, core for core, against the biggest baddest hot running desktop 125W i9 10900K, 105W AMD Ryzen 9 5950X etc.

It's got fewer cores, of course.

So in the same sense that a 4 cylinder engine is just as powerful as a V-8, cylinder per cylinder?

That comparison would make a lot more sense if cars ran on 1 cylinder most of the time.

My 4-cylinder car with turbo embarrasses most V-8s ever made, while using much less fuel on average.

[SilverSolder]

I thought this thread (including the Anandtech link earlier) has established that the new Apple SoC is primarily efficient, rather than being the ultimate in absolute performance - i.e. the Intel and AMD processors it is being compared with are able to draw much more power if needed, at which point they blow the new Apple chip away? 

Nope.

The M1 is right up there, core for core, against the biggest baddest hot running desktop 125W i9 10900K, 105W AMD Ryzen 9 5950X etc.

It's got fewer cores, of course.

So in the same sense that a 4 cylinder engine is just as powerful as a V-8, cylinder per cylinder?

That comparison would make a lot more sense if cars ran on 1 cylinder most of the time.

My 4-cylinder car with turbo embarrasses most V-8s ever made, while using much less fuel on average.

I can absolutely live with a 4 cylinder turbo instead of a naturally aspirated V8,  but if all else is equal (i.e. put a turbo on the V8 as well)...   the v8 will be twice as powerful.

I don't know about you, but the workload on my PC is "lumpy"...  long periods of light load, and then suddenly a massively intensive task that the user (me!) wants completed by yesterday! 

If you are not the kind of user that ever causes intensive loads, then you don't need a V8 or a turbo 4...   you are a non-challenging use case.  I think that is the target market, initially, because most users fit that profile (most people don't have V8 or turbo 4 cars either).

[brucehoult]

I thought this thread (including the Anandtech link earlier) has established that the new Apple SoC is primarily efficient, rather than being the ultimate in absolute performance - i.e. the Intel and AMD processors it is being compared with are able to draw much more power if needed, at which point they blow the new Apple chip away? 

Nope.

The M1 is right up there, core for core, against the biggest baddest hot running desktop 125W i9 10900K, 105W AMD Ryzen 9 5950X etc.

It's got fewer cores, of course.

So in the same sense that a 4 cylinder engine is just as powerful as a V-8, cylinder per cylinder?

That comparison would make a lot more sense if cars ran on 1 cylinder most of the time.

My 4-cylinder car with turbo embarrasses most V-8s ever made, while using much less fuel on average.

I can absolutely live with a 4 cylinder turbo instead of a naturally aspirated V8,  but if all else is equal (i.e. put a turbo on the V8 as well)...   the v8 will be twice as powerful.

I don't know about you, but the workload on my PC is "lumpy"...  long periods of light load, and then suddenly a massively intensive task that the user (me!) wants completed by yesterday! 

If you are not the kind of user that ever causes intensive loads, then you don't need a V8 or a turbo 4...   you are a non-challenging use case.  I think that is the target market, initially, because most users fit that profile (most people don't have V8 or turbo 4 cars either).

For the intensive massively parallel loads, I have a 32 core Threadripper.

For anything that uses four threads or fewer my M1 Mac Mini is faster than the Threadripper.

I've updated my trivial little primes numbers benchmark results to include the M1, both native and using Rosetta. It kills on this task.

http://hoult.org/primes.txt

Code: [Select]

//     2.735 sec i7 8650U @ 4.2 GHz               242 bytes  11.5 billion clocks
//     2.795 sec Mac Mini M1 @ 3.2 GHz (?)        212 bytes   8.9 billion clocks
//     2.872 sec i7 6700K @ 4.2 GHz               240 bytes  12.1 billion clocks
//     2.925 sec Mac Mini M1 @ 3.2 GHz Rosetta    208 bytes   9.4 billion clocks
//     3.448 sec Ryzen 5 4500U @ 4.0 GHz          242 bytes  13.8 billion clocks
//     3.515 sec Threadripper 2990WX @ 4.2 GHz    242 bytes  14.8 billion clocks

I have other larger tests in progress, such as building a RISC-V cross-compile toolchain with bunutils, gcc, gdb, newlib. I'll report back soon.

[bd139]

Interesting comparison....

Ryzen 3600 / gcc 9.3.0:

Code: [Select]

3713160 primes found in 3238 ms
276 bytes of code in countPrimes()

Ryzen 3600 / clang 10.0.0

Code: [Select]

3713160 primes found in 3517 ms
240 bytes of code in countPrimes()

DISCLAIMER: these are run inside virtualbox not on native hardware.

Will have native hardware and a 5900X to compare to if anyone ever get some fucking stock in 

Edit: slightly disappointed in clang if I'm honest.

Edit 2: Xeon Platinum 8151 @ 3.4GHz (AWS z1d) / CentOS 7 gcc 7.3.1

Code: [Select]

3713160 primes found in 3505 ms
244 bytes of code in countPrimes()

Edit 3: "AArch64 Processor [413fd0c1]" AWS g6g graviton 2 / AWS Linux gcc 7.3.1 (think its centos recompiled)

Code: [Select]

3713160 primes found in 6377 ms
276 bytes of code in countPrimes()

Slooow.

M1 = exceptionally good.

[SilverSolder]

[...]

For the intensive massively parallel loads, I have a 32 core Threadripper.

For anything that uses four threads or fewer my M1 Mac Mini is faster than the Threadripper.

[...]

I don't normally switch between different PCs depending on the load...   I have one workstation, it does many jobs at the same time by running many VMs, and therefore benefits from many cores.  This is probably a heavier than average workload, but that's why processors with more cores exist.

I also use applications like video editors and others that can make use of parallel processing - even if this is not used all the time, it is nice that it is available when it is.

Your benchmark is interesting and definitely shows the M1 core itself is no slouch, in the right circumstances.