M1 Macs Reviewed

[leadeater]

1 hour ago, LAwLz said:

We don't know, but right now the most probable answer is that eGPU support can be added through software updates. 

One possible showstopper right now might be drivers. We already know that kernel drivers are not compatible with rosetta2 so Apple can't just load the old drivers. I can't imagine AMD or Nvidia being that willing to develop ARM based drivers for Apple either. 

"Hey, we know that we just gave you the middle finger and developed our own GPU, but can you please spend time and money developing drivers for our new product that is competing with you?". 

I'm honestly not sure of the point, the GPU in the M1 is actually fairly powerful and has a lot of benefits of being within the SoC and unified memory. From what I have seen the M1 GPU in productivity/professional applications is as fast as or faster than most or any dGPU options of existing Macs so adding in eGPU support is of no great benefit unless you need very strong GPU compute or gaming.

 

Far as I see it Apple is EOL eGPU entirely, dGPU performance efficiency within applications is going to have to a hell of a lot more efficient to stick around on the Mac OS platform. A bigger SoC with 4 channel memory and more GPU cores will just tip that even further. You don't need a dGPU if it's actually going to be slower.

[Video Beagle]

18 hours ago, Bombastinator said:

It’s got some hefty weaknesses.  It’s gpu is great for an iGPU, but iGPUs still suck, and there is NO WAY to stick on anything else.  There’s also a hard memory cap.

 

I'd imagine such things will change with the M2 or M1 Pro or whatever they're naming scheme will be for future things, no? Like, (and I repeat what I've said often, this is "dragons be here" territory of the computer for me), you'd have the onboard memory of the SOC, then a "secondary" made of dimms?

 

4 hours ago, Carstenpxi said:

P.S.  A consequence of the tighter system design described above is the gradual fading away of hobbyist “tinkering” industry. 

Yeah.. it's literally what happened with the car industry... is it that folks in the forum and similar are in a bubble to not get this is how technology goes? and as @Carstenpximentioned the popularity of rasberry pi etc is similar to the car kit movement.

 

 

1 hour ago, LAwLz said:

Well Linus has already made his mind up about the M1 so he has to find some reason to hate it. 

 

Edit: after watching the short circuit video I think he sounded a bit less biased than his previous videos. 

He is still very much a pessimist about it though. 

 I don't hear pessimism...I hear bitterness.  Little asides sniping at others who got test units a week ago, ignoring that he gets test units of stuff others don't, also ignoring that he sounds whiny and unprofessional, both in terms of his continued non-importance to Apple and to his relationship with the reviewers..his peers... who did get early access, some he's friends, or at least friendly, with, with an implication that they're not as "good" as him, so they must suck up to Apple to get where they are.

 

I would really hope someone pulls him back, because it's not a good look for him or the channels, but frankly, I think mostly he has a support system of yes men to ensure him he's right.

[Bombastinator]

5 hours ago, avg123 said:

exactly. and since he is being rebutted by the tech community because of his stupid video, he now has to stick to his opinion because of his big ego.

The opinion appears to be he didn’t like the Apple presentation, which is something that has been echoed by several reviewers. 

[Bombastinator]

5 hours ago, LAwLz said:

Apple did have eGPU support before but didn't market it, so I don't know where you're getting this idea from. 

 

As for the "if you don't like it don't watch it" argument:

 

Video seems more like 2 arguments rather than 6, and the points are still interconnected.  Points are reasonable though.  Egpus aren’t the only thing thunderbolt is needed for.  There are several types of peripherals that need thunderbolt. 
 

As for egpus specifically, thunderbolt egpus were used a fair bit by Mac users still trying to do work on macs rather than switching to microsoft mostly because Apple stubbornly refused to put high end GPUs in their products, and there are people who want to use macs AND use powerful GPUs.  They will put high end MOBILE GPUs in their products but someone at Apple apparently just doesn’t get that just because a mobile gpu has the number of a powerful card doesn’t mean it has the POWER of that powerful card.  With pro uses it’s not about hardware names it’s about render time.  Highly paid professionals are highly paid, and if one of them is sitting around twiddling their thumbs waiting for a render literally everyone hates it.  The company owner hates it, the operator hates it, the client hates it, everyone. 
There are types of pro level work where THE most amount of memory possible and THE fastest gpu that can be had still aren’t nearly fast enough.  It’s merely that that is all that can be had.  Sure you could pay $4k for a pro iMac that had dual mGPUs, but those dual mGPUs were STILL slower than a single mid grade desktop GPU.  That such a machine also cost a third as much is kind of irrelevant.  Nice, but not the main point.  Users would run into the problem of people screaming at them “why the HELL aren’t you using a machine fast enough to do the work?!  They EXIST! You just have to stop using a Mac”

[avg123]

41 minutes ago, Bombastinator said:

The opinion appears to be he didn’t like the Apple presentation, which is something that has been echoed by several reviewers. 

nobody else made a whole video whining about it.  he made fool out of himself because every other reviewer had review units and knew apple was right and linus was wrong.

[BigDamn]

18 hours ago, LAwLz said:

Hey everyone! Look at how bad Apple is! 

Their quad core running at 10 watts can't even keep up with this 54 watt octa core! What a fail, am I right? 

 

This is totally evidence of geekbench being a bad benchmark as well since clearly this other benchmark gives a different result (except they don't, I just don't know the difference between single and multi core scores)! Everyone knows that if two benchmarks give different results then only the one that shows the result I want is valid! 

Geekbench has the M1 chip right behind the Xeon W-2140B in the 2017 iMac Pro in MULTI CORE. Now you tell me, would you rather have the M1, or the significantly faster Xeon?

 

It appears that it may be you who can't read benchmarks, because not only did you fail to realize that Geekbench has the M1 scoring closely to beefier chips in multi core, but you failed to realize that my graph shows more than just 8 core competitors. The M1 was beaten by ALL AMD mobile chips, 8 cores or lower. Remember that the M1 is technically an octa-core, with 4 powerful cores and 4 lower performance cores (which are utilized in heavy workloads), so this chip acts more like a 5-6 core CPU and loses to 6 core parts. This was hitting people with a dose of reality, rather than making a mockery of Apple.

[Bombastinator]

1 minute ago, avg123 said:

nobody else made a whole video whining about it.  he made fool out of himself because every other reviewer had review units and knew apple was right and linus was wrong.

The key there being “whole video”. There were other reviewers that echoed the same sentiment about the presentation.

[avg123]

1 minute ago, Bombastinator said:

The key there being “whole video”. There were other reviewers that echoed the same sentiment about the presentation.

nobody did. show me an example.

[saltycaramel]

3 minutes ago, BigDamn said:

Geekbench has the M1 chip right behind the Xeon W-2140B in the 2017 iMac Pro in MULTI CORE. Now you tell me, would you rather have the M1, or the significantly faster Xeon?

 

It appears that it may be you who can't read benchmarks, because not only did you fail to realize that Geekbench has the M1 scoring closely to beefier chips in multi core, but you failed to realize that my graph shows more than just 8 core competitors. The M1 was beaten by ALL AMD mobile chips, 8 cores or lower. Remember that the M1 is technically an octa-core, with 4 powerful cores and 4 lower performance cores (which are utilized in heavy workloads), so this chip acts more like a 5-6 core CPU and loses to 6 core parts. This was hitting people with a dose of reality, rather than making a mockery of Apple.

 

now explain that again in performance-per-watt terms or you’re basically saying nothing (raw core count? what? be it 10W or 100W?)

[Bombastinator]

1 minute ago, avg123 said:

nobody did. show me an example.

Sure.  

 

[Lord Vile]

2 minutes ago, BigDamn said:

Geekbench has the M1 chip right behind the Xeon W-2140B in the 2017 iMac Pro in MULTI CORE. Now you tell me, would you rather have the M1, or the significantly faster Xeon?

 

It appears that it may be you who can't read benchmarks, because not only did you fail to realize that Geekbench has the M1 scoring closely to beefier chips in multi core, but you failed to realize that my graph shows more than just 8 core competitors. The M1 was beaten by ALL AMD mobile chips, 8 cores or lower. Remember that the M1 is technically an octa-core, with 4 powerful cores and 4 lower performance cores (which are utilized in heavy workloads), so this chip acts more like a 5-6 core CPU and loses to 6 core parts. This was hitting people with a dose of reality, rather than making a mockery of Apple.

The M1 is currently inside 2 13" thin and lights, one of which has no active cooling and a SFF PC. The W2140B s a 120W chip with 16 threads. What did you expect? It beats everything in single core though. It also only loses to the 4900HS which is AMDs top mobile chip with 16 threads by about 10% when it's configured to use the full power nd a lot of results have the M1 levelling or beating the 4900HS.

[Lord Vile]

2 minutes ago, saltycaramel said:

 

now explain that again in performance-per-watt terms or you’re basically saying nothing (raw core count? what? be it 10W or 100W?)

The Xeon is a 120W part and the 4900HS which is AMDs top part with 8 cores and 16 threads at 35W beats the M1 by 10% sometimes. Sometimes it loses depends on the implementation. Would love to see a comparison of the M1 in a MBP on battery vs the 4900HS on battery and see how they score and how long the battery lasts when the chips are stressed.

[avg123]

5 minutes ago, Bombastinator said:

Sure.  

 

this video is from after 

[saltycaramel]

1 minute ago, Lord Vile said:

The Xeon is a 120W part and the 4900HS which is AMDs top part with 8 cores and 16 threads at 35W beats the M1 by 10% sometimes. Sometimes it loses depends on the implementation. Would love to see a comparison of the M1 in a MBP on battery vs the 4900HS on battery and see how they score and how long the battery lasts when the chips are stressed.

 

Can’t wait to see ~30W 12-core (8p + 4e) M1X chips in the all newly designed 14” and 16” MBPs...they’re gonna annihilate a lot of desktop CPUs and any laptop CPU under the sun..

 

T minus 120-180 days

[Bombastinator]

9 minutes ago, avg123 said:

this video is from after 

So your complaint is that he said it first?

[Lord Vile]

8 minutes ago, saltycaramel said:

 

Can’t wait to see ~30W 12-core (8p + 4e) M1X chips in the all newly designed 14” and 16” MBPs...they’re gonna annihilate a lot of desktop CPUs and any laptop CPU under the sun..

 

T minus 120-180 days

Be interesting if the release larger chips, I'm happy with my 16" Pro but in a couple of years a new ARM 16" might get me to upgrade.

[saltycaramel]

Oh, they will. Quanta is preparing for mass production of the 16” as we speak.

 

Have Intel and AMD got anything ready in the next 120-180 days to counter this?

 

This is what we should be talking about.

 

That’s the real story, not only these low end appetizers M1 Macs. (or “3 iPads”, as one youtuber eager to become the next Steve_Ballmer_laughs_at_the_iPhone meme once called them)

 

The clock is ticking.

 

Like Thanos, it’s inevitable at this point.

[BigDamn]

24 minutes ago, Lord Vile said:

The M1 is currently inside 2 13" thin and lights, one of which has no active cooling and a SFF PC. The W2140B s a 120W chip with 16 threads. What did you expect? It beats everything in single core though. It also only loses to the 4900HS which is AMDs top mobile chip with 16 threads by about 10% when it's configured to use the full power nd a lot of results have the M1 levelling or beating the 4900HS.

You missed the point. I was explaining that Geekbench has the M1 right behind the W2140B in multi-score, which is why Geekbench isn't a good benchmark for comparisons.

[Lord Vile]

Just now, BigDamn said:

You missed the point. I was explaining that Geekbench has the M1 right behind the W2140B in multi-score, which is why Geekbench isn't a good benchmark for comparisons.

Why? It seems to be pretty accurate in terms of performance.

[Lord Vile]

18 minutes ago, Bombastinator said:

So your complaint is that he said it first?

My issue is the "3 iPads" think. He wouldn't call an i7 K SKU a tablet chip because a surface runs a core M3 so why call the new Macs iPads. Personally think Linus is going to get ridiculed when they try and make the M1 Macs look bad by trying to bench them like a gaming PC.

[BigDamn]

2 minutes ago, Lord Vile said:

Why? It seems to be pretty accurate in terms of performance.

The W2140B is significantly faster than the M1 in multicore. As you previously noted, that's not surprising.

[Lord Vile]

2 minutes ago, BigDamn said:

The W2140B is significantly faster than the M1 in multicore. As you previously noted, that's not surprising.

Depends on the workload and application, if the program only scales to say 8 cores i'd expect the Mac to take it. Also with the new systems you're likely to be using Logic or FCP which will take advantage of the neural engine rather than jus compute cores meaning the usual render benchmarks etc with be fairly pointless.

[Bombastinator]

4 minutes ago, Lord Vile said:

My issue is the "3 iPads" think. He wouldn't call an i7 K SKU a tablet chip because a surface runs a core M3 so why call the new Macs iPads. Personally think Linus is going to get ridiculed when they try and make the M1 Macs look bad by trying to bench them like a gaming PC.

I have no doubt.  Still needs to be done though. 

[saltycaramel]

2 minutes ago, BigDamn said:

The W2140B is significantly faster than the M1 in multicore.

Says who. 
Doing what.

That’s the whole conundrum.

All bets are off depending on how these are used. (anandtech went in-depth about the low level differences between the CPUs)

[Lord Vile]

1 minute ago, Bombastinator said:

I have no doubt.  Still needs to be done though. 

Yup running every video editor available but Final cut is the best way to bench a Mac because that's why people buy Macs. To avoid all of the lovely optimised exclusive software that performs better than anything else on equivalent hardware. Oh and lets bench everything on wall power too because apple totally doesn't approach a laptop like a laptop and they definitely don't optimise the hardware so it can perform flat out on battery if you can't get to a socket or are on a plane where you can't draw 200W from the wall.