New Mac Mini + WOW thought..?

[Mouri Kogorou]

does that mean the new ipad air will be able to run wow with the native support to ARM processors?

[Chakah]

does that mean the new ipad air will be able to run wow with the native support to ARM processors?

No. iOS can't run MacOS Apps...

(yet?) Nah... that would be crazy!

[Gaidax]

does that mean the new ipad air will be able to run wow with the native support to ARM processors?

Certainly one step closer.

I bet eventually WoW will be running on iPads at least in some capacity.

[Thunderball]

Citation needed.

Machines are just now getting into peoples hands, so benchmarks will be forthcoming over the next days/weeks.
Here are a couple of other early ones:

https://www.macrumors.com/2020/11/16...sd-benchmarks/
https://www.macrumors.com/2020/11/16...nch-benchmark/

I quoted you. You're showing mobile silicon to be competitive with desktop CPUs in mobile benchmarks. M1 is not going to be any different. Actual performance, however, wont show any of that.

[Temp name]

https://us.forums.blizzard.com/en/wow/t/mac-support-update-november-16/722775

M1/ARM is supported.


LOL that didn't age well.

I'm surprised they got it that fast.. I expected a month or two delay

[Kagthul]

Are you really comparing platforms that work on different instruction sets. If the benchmark were written using the actual x86-64 instruction set the a14 literally wouldn't be able to run it...

Thats what Rosetta 2 is for. Not that there isn't performance loss running stuff via Rosetta.

You realize that's what the R in Risc which is the R in Arm means?

None if this makes sense.

I wouldn't go that far.

Arm is used for low power devices like phones and tablets because of its efficiency. Low power implies better thermals. But not better performance. It makes sense for the the form factors they released so far, I guess. A small laptop and a small form factor pc. But there's nothing "pro" about this. They have the advantage with process node, but again that drives better efficiency, not performance.

Again, not entirely. RISC can be high performance. And the benefit of RISC, as Apple touted long ago, is that with proper support a 2.0Ghz PowerPC G5 was able to keep pace with a 3-ish Ghz P4.

Problem is that that tends to fall off eventually, as RISC chips in the past have had trouble breaking high clock speeds, for whatever reason (im simply not that in the know on the physical/physics differences between CISC and RISC), but it was the primary reason Apple ditched PowerPC for Intel - The G5 could keep up with "faster" Intel Pentium chips, but IBM just couldn't get PPC past around 2.5ghz. And they DEFINITELY couldn't get it into a power-efficient package for laptops.

Well likely see the same with Apple Silicon/ARM-based RISC chips as well, eventually. A decent amount of clock speed deficit can be made up by being more efficient (if an application has to implement X instructions instead of X*5, the clock advantage of the CISC chip is largely offset) but that actually requires quite a bit MORE programming and work to really optimize for RISC.

You could probably get better performance and performance per watt from a nehalem i7.

No. That benchmark is, AFAIK, showing Geekbench running under Rosetta 2 - meaning its actually posting a lower score than it otherwise would if it were native. Meaning that M1 is beating modern X86 chips at least at single core.

Maybe. I could be wrong, and that M1 Mac may be running the ARM version of Geekbench, which does exist, and that WOULD make that graph useless.... since one is optimized for ARM and one is optimized for RISC, the results arent actually comparable.

The best way to bench the M1 against an X86 wouldn't be (an already shitty, and nearly useless; Geekbench is pretty trash) synthetic benchmark - i would be to subject them to tests with measurable metrics that dont produce some arbitrary synthetic number....

Like, say Cinebench. Which is something you can measure simply by "how fast did it complete the render?"

Something like, oh, i dunnoo......

This: https://hexus.net/tech/news/cpu/1468...ores-revealed/

Notice that this is an ARM native version of Cinebench (which is, itself, just a canned run using the EXACT same engine/software as the commercial product, Cinema 4D).

Notice that the M1 isn't showing real-world performance anywhere near what the Geekbench result would imply.

It is NOT faster Single-core than modern X86-64 CPUs in most cases (the numbers listed in the graph are actually for the A12Z/X that was in the dev mini, and the M1 is closer to ~1400 single core, which would, quite honestly, be solely due to faster clocks than the A12Z which was about 900mhz slower).

Its surpassing laptop parts pretty handily (about ~300 points faster), but its barely hanging in with a bone-stock 3600X. Given that a 5600X is substantially better than a 3600... the M1 isn't exactly lighting bedsheets on fire.

And given that that's STOCK operation, and a 10600K overclocked beats the 3600X by a decent bit.... a 5ghz+ Intel CPU is going to beat it pretty handily.

Again, as ive been saying all along.... its not a bad chip. Its perfectly OK for daily driving, and itll be a great laptop chip since it sips power.

But it isn't all Hookers and Sunshine.

(*note: before you go on with "but Cinebench scores are just points too, so how is it better/more accurate than Geekbench?!?!" - well, for one thing Geekbench is coded like ass, and for another, the number in Cinebench is just generated from how long it took to complete the render).

[Chakah]

I quoted you. You're showing mobile silicon to be competitive with desktop CPUs in mobile benchmarks. M1 is not going to be any different. Actual performance, however, wont show any of that.

Nah..

Geekbench 5 isn't a mobile only benchmark. and SSD speed contributes to your hypothetical 'general purpose workload'.

Citation still needed...

[Kagthul]

Nah..

Geekbench 5 isn't a mobile only benchmark.

It basically is, actually. The different versions dont produce comparable numbers.

You can really only compare Geekbench results for the individual architecture... so, comparing an Exynos chip to an A12X or what have you.

When you try to compare them cross-architecture the numbers never add up right because Geekbench isn't actually using some cross-architecture/platform comparable metric - like, say, Cinebench does (because the score is generated by a simple equation based around how long the render actually took).

Which is why if you follow the link i posted above (to Cinebench results) youll see that while the M1 does quite well (beating laptop parts of similar power consumption handily, and comparable to a generation+ old desktop part running bone-stock), but it isn't the Second Coming.

Its an 8 core part that puts up numbers like a 6-core part, which is still impressive because there's no SMT/HT, but it isn't "beating" Ryzen 5000 or CometLake-S like the Geekbench scores listed earlier seem to show. Its roughly equivalent to older parts, but current gen parts are ~10-20% faster (especially depending on RAM on Ryzen) than that... so...

Like i said.

Its good.

But it isn't all that and a bag of chips.

[Thunderball]

SSD speed contributes to your hypothetical 'general purpose workload'.

Everything's clear with you. Have a good day.

[Chakah]

Which is why if you follow the link i posted above (to Cinebench results) youll see that while the M1 does quite well (beating laptop parts of similar power consumption handily, and comparable to a generation+ old desktop part running bone-stock), but it isn't the Second Coming.

Its an 8 core part that puts up numbers like a 6-core part, which is still impressive because there's no SMT/HT, but it isn't "beating" Ryzen 5000 or CometLake-S like the Geekbench scores listed earlier seem to show. Its roughly equivalent to older parts, but current gen parts are ~10-20% faster (especially depending on RAM on Ryzen) than that... so...

Like i said.

Its good.

But it isn't all that and a bag of chips.

From your article:
"M1 Cinebench R23 scores:1,498"
AMD Ryzen 5 3600X was 1300
AMD Ryzen 7 4800HS was 1230

so... ya... it is beating a Ryzen for Cinebench 23 single core.

[ldev]

Instead of hating apple, buy their stock people. At least you will get a bit richer while Apple will be making your ass hurt.

- - - Updated - - -

From your article:
"M1 Cinebench R23 scores:1,498"
AMD Ryzen 5 3600X was 1300
AMD Ryzen 7 4800HS was 1230

so... ya... it is beating a Ryzen for Cinebench 23 single core.

They're trying to say that literally the lowest of the low end first gen only 4 high perf core CPU that's put in literally the cheapest sub 1 thousand FANLESS notebook is still not going to beat 12 or 16 or 24 core Intel/AMD chips in overall multitasking

Yea, they are that stupid.

[Kagthul]

From your article:
"M1 Cinebench R23 scores:1,498"
AMD Ryzen 5 3600X was 1300
AMD Ryzen 7 4800HS was 1230

so... ya... it is beating a Ryzen for Cinebench 23 single core.

Yes, its beating a generation and a half old Ryzen.

It is not beating a 5600X.

Reading is pretty hard for you, i can see.

[Gaidax]

Yes, its beating a generation and a half old Ryzen.

It is not beating a 5600X.

Reading is pretty hard for you, i can see.

It is pretty remarkable that something like M1 - a fanless SoC that is present in MacBook Air - manages to beat a full out desktop midrange+ CPU even of generation and half in a single core performance and at a fraction of power at that.

Can you imagine? They have this thing practically in latest iPad Pros. I can only wonder what M2 will be then.

- - - Updated - - -

People with their senseless hate for platform/company can't even grasp just what they have accomplished there.

[Kagthul]

Instead of hating apple, buy their stock people. At least you will get a bit richer while Apple will be making your ass hurt.

I know you're barely able to read and all...but... i use a Mac. And an iPad. And ive always used Apple. Since i was like 8 years old with the original Beige Toaster box.

I dont hate Apple. Im just trying to inject some fucking sanity. Which isn't your strong suit, i get it.

They're trying to say that literally the lowest of the low end first gen only 4 high perf core CPU that's put in literally the cheapest sub 1 thousand FANLESS notebook is still not going to beat 12 or 16 or 24 core Intel/AMD chips in overall multitasking

Yea, they are that stupid.

Might want to look in the mirror when you use the word "stupid".

To address the highlighted: its also the EXACT SAME CPU that they put in the MacBook PRO 13" (which has active cooling) and the Mini (which has active cooling)... so, i dunno, you might want to keep up.

I was focused solely on single-core performance.

People were trying to say it beats modern X86-64 chips.

It doesn't. It beats generation and a half old X86-64 chips. And not by a lot, considering its being measured bone-stock with DDR4-2400. Geekbench showed it beating a 5950X.... Which it most certainly DOES NOT if it can barely beat a 3600.

Its a good chip. But it isnt, even remotely, the "first gen" - Its the exact same cores as the A14 without thermal limitations and higher clock speeds. (And more cores) So its like.. their 8 or 9th gen chip ( i honestly dont remember what # they started on when they started using the A-series designator)

And performance chip generation to chip generation is ALREADY falling off. The M1 is barely faster (only about 9%) than the A13. The A13 was only about a 15% uptick from the A12 and was barely faster than the A12X at all (because the A12X was in larger devices with less thermal constraints)

There's absolutely no guarantee that a titular M2 will be massively faster than the M1. If we go by the last 3 years of Apple chips... another 10 or 15%. Which will bring it up to par with Zen 3 (Ryzen 5000) and Comet Lake-S. Problem is early next year we're going to get Rocket Lake S (with a 20% IPC uplift), and mid next year well see Zen 4 revealed if not released - and early early leaks are showing Zen 4 having a significant uplift from Zen 3.

Its good. But it isn't all that and a bag of chips. It isn't all hookers and sunshine, and if they cant produce massively multicore performance (no evidence of that either way, but typically ARM derived chips with lots of cores have SUPER low clock speeds) then its a complete non-starter in the Pro scene. And the "its all soldered down, buy a new machine, pleb" any time you might want to upgrade is a giant shit sandwich.

Notice:

https://www.cpu-monkey.com/en/cpu_be...single_core-15

it doesn't even beat a 4 core/8 thread, low power Tiger Lake laptop part. It comes close... but doesn't beat it. (And Rocket Lake-S is shaping up to be 10-15% faster than the Tiger Lake parts + ~1.2ghz faster clocks).

It doesn't beat any of the new Ryzen chips. (And all of those, again, are being measured bone-stock with base-spec DDR4 RAM; especially with Ryzen 5000, good RAM can get you another 10% performance).

So, again:

Its a good chip. Its not the second coming.

[ldev]

To address the highlighted: its also the EXACT SAME CPU that they put in the MacBook PRO 13" (which has active cooling) and the Mini (which has active cooling)... so, i dunno, you might want to keep up.

Yes, the base model MBP. Not the higher end 13", not the 16" MBP.

Yes, the base model desktop. Not the iMac, not the iMac Pro, not then Mac Pro.

Keep up.

So, again:

Its a good chip. Its not the second coming.

What? I know. I have zero interest in M1 laptop. It's too low end for me. Waiting for RDNA2 mbp and imac for now and later we'll see how well high end apple silicon is doing. Keep up.

[kukkamies]

They're trying to say that literally the lowest of the low end first gen only 4 high perf core CPU that's put in literally the cheapest sub 1 thousand FANLESS notebook is still not going to beat 12 or 16 or 24 core Intel/AMD chips in overall multitasking

Yea, they are that stupid.

Does anyone care about single core past 100fps+ gaming? Not that much gaming in Macs anyway.
Nah, it's that it doesn't even compete with the lowest of the low 200usd desktop cpu. 1000$ laptop with the Ryzen 4800H still beats it by over 30% in multicore cinebenches. It is good, but the hype does look more overblown every day.

[Kagthul]

Does anyone care about single core past 100fps+ gaming? Not that much gaming in Macs anyway.

Right? The only thing where single-core performance really matters is the one thing that the M1 is still going to be terrible for (gaming) because of the iGPU being about as good as almost 3 generation old low-end parts.

Nah, it's that it doesn't even compete with the lowest of the low 200usd desktop cpu. 1000$ laptop with the Ryzen 4800H still beats it by over 30% in multicore cinebenches. It is good, but the hype does look more overblown every day.

Yup.

Multicore CB:

https://wccftech.com/intel-and-amd-x...hmark-results/

Relevant:

The problem is that ARM architecture doesn't clock very high and has high leakage at high power levels

Same thing happened with PowerPC, and other RISC designs. They simply dont clock that high. And high-end PowerPC parts for G5 desktops sucked down MASSIVE power for CPUs at the time and required bespoke liquid cooling for the dual-processor machines.

[epo75]

Kagthul

Not sure who you are, but what I know is your either a AMD fanboy or just a Apple hater. You write big stories but none of them are correct.

First of all, this is a chip with FOUR performance cores, and FOUR low power cores, it is the low-end of Apple chips. Comparing it with a full blown 8 cores chip is a joke.
Secondly this is a INTEGRATED gpu...
Thirdly, so what this chip can't clock to 5GHZ? They are scalable and can be put in clusters.

All other stuff your writing is crap too and not correct, anyways not gonna put more time in it.

AMD Zen 2 and 3 are great chips, but this M1 is amazing, it simply is and there is so much more to come. Also more to come from AMD which is great. Stop acting like a kid.

[ldev]

Holy shit, 6 core amd cpu barely destroys 4 high perf core low end apple cpu and 8 core amd cpu destroys 4 high perf core low end apple cpu, NOOOOOO.

Apple - from a trillion dollar company to bankrupt company in one night. Rekt.

Anyways, Apple is focused on selling experience. A very fucking nice experience. All the small things imaginable:
* instant wake up - you open the lid and it's on
* silence. So. much. silence. My new Mac mini just doesn't make a sound, and for the performance it delivers, it's fucking insane
* proper dpi scaling - start dragging apps window from non scaled screen to scaled one - and in scaled screen that part of apps window is scaled properly, on Windows - apps part that's in scaled window is super small, then it lags for a moment, fucking redraws itself and then it's big on non scaled screen, Jesus f christ
* bluetooth works - flipping from bluetooth headsets to external audio is instant and not a problem for any app, meanwhile on windows 1/3 of apps stop producing sound, 1/3 continue to play on bluetooth headsets, and 1/3 just crash.

I could go on, sadly didn't use Windows PC for some time now.

So even if Apple Silicon will be lower performance - most Apple users don't care, they will chose a nice leather massaging chair over a fucking wooden bench that has PERFORMANCE.

Done discussing Apple with tasteless common peasants - M1 adds to the posh experience and it's a fucking awesome CPU and Windows with AMD or Intel have literally nothing to offer besides cheap performance and can't hold a candle against M1. I'd rather daily drive S class than a fucking Ariel Atom.