Another ARM(1) - Apple event introduces new low-end Macs as well as updated OS

[Dracarris]

8 hours ago, leadeater said:

Good luck when it's both out of support and also there isn't an AASP in your town. If it's not in warranty and Apple would fix it you pay for the part and the time, neither being cheap. If I take myself out of the equation for doing it as that's not fair my local PC store is cheaper than what Apple charges.

 

1 hour ago, GDRRiley said:

is that a joke. I'll wait 2-4 weeks get charged 10x the price I should because they just play dart board with what is wrong.

HP, dell, lenovo can all get my parts in under 2 days time.

I have absolutely no clue what you guys are talking about. All the experience with repairs at Apple I have experienced either personally or with friends/family, go like this:

- You bring the device to the Apple store after they gave you an appointment

- They do the repair either same-day or you can pick the device up the next day

- With one exception, everything so far was either covered under warranty, an repair program or any cost was otherwise waived. The one exception would've been a logic board swap which I did not do at the end. Labor cost was very low and quite reasonable. Since the device in question was not bought directly at Apple but a reseller, I had to take it there for a warranty repair (AASP)

- The repair above was during the first Corona lockdown and both Apple and AASP were understaffed and everyone bringing in broken stuff due to the heavy use in Home Office. As a consequence repair times were vastly increased: Both Apple and the AASP sincereley apologized in advance that the repair is going to take 1 week under these special circumstances. The AASP indeed repaired my device within one week, free of charge of course.

 

I know hating at Apple online is en-vogue these days, but it simply is often far from reality.

 

1 hour ago, GDRRiley said:

I pulled open my probook 445R G6 and started looking at IC

  Reveal hidden contents

found a GST5009B-LF Ethernet IC https://www.kynix.com/Detail/520384/GST5009B-LF.html and  I can get pinouts for it as well

I can't flip the board over as the machine is still on so that was all the ICs on this side

 

That is a off-the-shelf magnetic coupler for ethernet, not at all custom. Simply not available easily from western distributors (mmouser, digikey). If you want to have it, you have to buy in bulk from Chinese sites.

[LAwLz]

Would it be against the forum rules to make a list of all the idiotic and stupid comments people have made in this thread so that I can go back in tell everyone "told you so" once we start getting a ton of benchmarks out there? Kind of like a "Wall of I-am-clueless-and-dont-understand-technology-but-comment-on-it-anyway"?

Or is that a waste of time because those people will just move the goal posts and/or pretend like they never said those stupid things, and/or start saying benchmarks doesn't matter?

 

It would be interesting to know how many of the people saying "Geekbench is a bad benchmark" even knows what Geekbench does and what it tests. I am fairly sure a majority of people saying it are saying it because they heard Linus Torvalds say he thought it was bad 7 years ago.

[Sauron]

34 minutes ago, LAwLz said:

Would it be against the forum rules to make a list of all the idiotic and stupid comments people have made in this thread so that I can go back in tell everyone "told you so" once we start getting a ton of benchmarks out there? Kind of like a "Wall of I-am-clueless-and-dont-understand-technology-but-comment-on-it-anyway"?

Or is that a waste of time because those people will just move the goal posts and/or pretend like they never said those stupid things, and/or start saying benchmarks doesn't matter?

 

It would be interesting to know how many of the people saying "Geekbench is a bad benchmark" even knows what Geekbench does and what it tests. I am fairly sure a majority of people saying it are saying it because they heard Linus Torvalds say he thought it was bad 7 years ago.

Damn, imagine getting so defensive when someone says we should probably wait for other benchmarks before jumping to conclusions.

[Helpful Tech Witch]

36 minutes ago, LAwLz said:

It would be interesting to know how many of the people saying "Geekbench is a bad benchmark" even knows what Geekbench does and what it tests. I am fairly sure a majority of people saying it are saying it because they heard Linus Torvalds say he thought it was bad 7 years ago.

or Linus Sebastian say it is bad 7 hours ago lol.

[LAwLz]

Just now, Sauron said:

Damn, imagine getting so defensive when someone says we should probably wait for other benchmarks before jumping to conclusions.

The only thing I am hearing is "I don't know how this benchmark works but I am going to say it is bad and doesn't matter because it shows results I don't like".

 

I wonder what people will say when results in Cinebench are good too. Maybe people will finally start saying Cinebench is a bad benchmark and doesn't matter?

 

 

But yes, I agree that we should wait for more benchmarks, mr:

Quote

Geekbench is meaningless for ARM vs x86 comparisons, it heavily favors the former.

Quote

Aggregate benchmarks in general are a bad way of measuring performance.

Quote

Apple themselves don't seem nearly as confident about this - they still sell 13" MBPs with Intel quad cores at a higher price. Why would they do this if the chip they put in their lowest end option was faster than their CURRENT top end 16" model?

Quote

We have high end ARM systems with plenty of cores and no power limitations and they don't blow out x86 chips with similar power designs in "real" benchmarks even though geekbench results for phones would suggest otherwise.

Quote

We'll have a much clearer picture once we get come cinebench results.

 

[LAwLz]

21 minutes ago, HelpfulTechWizard said:

or Linus Sebastian say it is bad 7 hours ago lol.

Linus Sebastian is a sales person and his technical knowledge is about as deep as a kiddie pool.

He knows how to run a benchmarking program in Windows and he knows about various features for consumer electronic products because the back of a box told him about it, or because a sales person from a manufacturer told him about it.

 

I've often said that building PCs as a job is basically the IT world's equivalence of flipping burgers at McDonalds. Pretty much everyone can do it since it requires very little skills or knowledge. It's boring. It pays very poorly. It's mostly done by teenagers.

If building PCs is the IT world's burger flipper at McDonalds, then Linus is the IT world's Ronald McDonald.

 

 

I haven't watched the WAN show yet so I don't know what Linus has said about the M1, but I am pretty sure he gets a ton of stuff wrong.

 

Edit:

Quote

07:35
Linus: so for example media encode and decode is handled via hardware blocks in Apple silicon, while previous Macs relied on software

This is just flat out wrong. Hardware decode and encode was handled in hardware blocks in Intel Macs too. I have made countless number of posts about this.

This is a very embarrassing mistake to make because it shows ignorance on so many levels. Here is a post from 2016 where I talk about it briefly:

  

On 11/7/2016 at 7:24 PM, LAwLz said:

Quote

Maybe you might be talking about Intel QuickSync in specific, but you don't have any idea what FCPX actually has that improves the quality while maintaing the speed

It is QuickSync, and it's not magic.

It's not a coincident that all devices that supports QuickSync, such as the dual core 13" Macbook Pro outperforms the 8 core Mac Pro when exporting from Final Cut. This benchmark should also disprove everyone who thinks it is because of optimizations because "OS X is great because it runs on specific hardware and that allows Apple to optimize it really well". If that was the case then the Mac Pro 2013 would destroy the MacBook Pro, which it doesn't.

 

 

Quote

07:35
Linus: so for example media encode and decode is handled via hardware blocks in Apple silicon, while previous Macs relied on software

I do, because it is. As you can see in the benchmark above, because of the media engine the dual core Macbook Pro outperforms the 8 core Mac Pro in this particular test. Remove that piece of hardware and the Macbook Pro would be terrible slow and the Mac Pro would run circles around it.

 

[Helpful Tech Witch]

3 minutes ago, LAwLz said:

I haven't watched the WAN show yet so I don't know what Linus has said about the M1, but I am pretty sure he gets a ton of stuff wrong.

He basically said that geekbench is too short to give a full picture. (which it is) Some run can happen when the cpu is still boosting, and makes something look a lot better than it is.

 

5 minutes ago, LAwLz said:

It's mostly done by teenagers.

If building PCs is the IT world's burger flipper at McDonalds, then Linus is the IT world's Ronald McDonald.

I disagree, Ronald McDonald doesn't now how to wire rgb, which is a skill that not many have.

[Sauron]

2 minutes ago, LAwLz said:

But yes, I agree that we should wait for more benchmarks, mr:

Yeah, I don't see anything in my posts that amounts to a gotcha. Maybe you should try to look at other people's posts with a little more good faith.

 

I don't trust aggregate benchmarks for the reasons I detailed; I said we should wait for more results, specifically for cinebench which is a single application benchmark that has an excellent track record and isn't only used because it's the only well known cross platform benchmark; I said Apple's behavior doesn't really track with a company that has just made a cpu that beats out its own highest end machinesm, while of course that's just a reason to be suspicious and not a guarantee of anything.

6 minutes ago, LAwLz said:

The only thing I am hearing is "I don't know how this benchmark works but I am going to say it is bad and doesn't matter because it shows results I don't like".

Then maybe start listening more? Geekbench gets routinely blasted every time it gets brought up, definitely not just in this case or just for Apple. People are saying they don't trust it because it has a bad track record and it works in a way that often leads to misleading results particularly across architectural barriers, be it due to aggregating a bunch of scores that aren't always closely related or as others brought up because it doesn't really stress a cpu's thermal limits (especially considering Apple is known to push their cpus to insane levels for a short time and then throttle to hell); this doesn't automatically mean it's always wrong, it just means that it's not reliable. It also doesn't always directly correlate with well known application specific benchmarks. This is why people are saying we should wait - not that it's definitely wrong, just that we. should. wait.

15 minutes ago, LAwLz said:

I wonder what people will say when results in Cinebench are good too. Maybe people will finally start saying Cinebench is a bad benchmark and doesn't matter?

Cinebench reflects performance in a specific application and is known to track pretty well with other intensive workloads. Because it's actually doing something and not just running a bunch of synthetic math that's supposed to simulate that something it has a real world meaning on what performance you can expect it to offer. Of course, just one benchmark would still not be enough to show you the whole picture but I'm inclined to trust it a lot more than geekbench.

[LAwLz]

3 minutes ago, HelpfulTechWizard said:

He basically said that geekbench is too short to give a full picture. (which it is) Some run can happen when the cpu is still boosting, and makes something look a lot better than it is.

That is true. 

It's worth noting that Apple's ARM processors does not have a boosting mechanic, so the point Linus brings up is an advantage to x86 and a disadvantage to ARM.

[piemadd]

Remember kiddos those cores aren't hyperthreaded. Also no boosting along with the fact that even if single core performance is better, more and more applications are multithreaded these days so we're going to have to see.

[LAwLz]

25 minutes ago, Sauron said:

I don't trust aggregate benchmarks for the reasons I detailed; I said we should wait for more results, specifically for cinebench which is a single application benchmark that has an excellent track record and isn't only used because it's the only well known cross platform benchmark;

I am also interested in Cinebench results. I wouldn't be surprised if they fairly closely mirrors the results from Geekbench though.

I don't understand what you mean by Cinebench having a good track record though. People like it because it is simple to run and provides both single threaded and multi threaded results. People don't like it because it is a real world benchmark or anything like that. I would say Geekbench is a more "real world" test than Cinebech is, because the programs in Geekbench are more widely used than the program Cinebench is based on.

 

 

25 minutes ago, Sauron said:

I said Apple's behavior doesn't really track with a company that has just made a cpu that beats out its own highest end machinesm, while of course that's just a reason to be suspicious and not a guarantee of anything.

And I think it does. I think it makes perfect sense.

 

 

25 minutes ago, Sauron said:

Then maybe start listening more? Geekbench gets routinely blasted every time it gets brought up, definitely not just in this case or just for Apple. 

Geekbench gets routinely blasted by people who doesn't have the technical understanding necessary to actually make the claims they make.

It's like saying "vaccines gets blasted on Facebook whenever they get brought up". Just because some loud and vocal minority who doesn't understand something says something is bad doesn't mean it is bad.

It's important to understand what benchmarks does to analyze the data they provide.

 

 

25 minutes ago, Sauron said:

People are saying they don't trust it because it has a bad track record and it works in a way that often leads to misleading results particularly across architectural barriers

[Citation Needed] on it giving misleading results across architectural barriers.

 

25 minutes ago, Sauron said:

be it due to aggregating a bunch of scores that aren't always closely related

That's what a general purpose benchmarking suite is suppose to do. Test a wide range of things and aggregate them into a generalized score. That's pretty much the only way to generalize the performance of a general purpose CPU. It doesn't make sense to do it any other way if you want a general overview of the performance of a processor. You have to do a wide range of tests since a CPU is purposely designed to do a wide range of tasks.

 

 

25 minutes ago, Sauron said:

or as others brought up because it doesn't really stress a cpu's thermal limits (especially considering Apple is known to push their cpus to insane levels for a short time and then throttle to hell)

[Citation Needed] on Apple pushing their ARM CPUs to the absolute maximum thermal limit.

That might be true for their x86 products, but I have seen no indication of Apple pushing their ARM CPU cores to the max in terms of thermals.

Also, like I said earlier, Apple's ARM chips does not have any boosting capability. The fact that the tests are short is a DRAWBACK when testing these chips against x86 processors.

x86 processors are designed to boost when they detect this type of task. They are designed to throw thermal management and power limits out the window and just go for maximum performance. The M1 can not do that.

Short tests = good for x86, bad for the M1.

Your understanding of this is completely wrong. You're saying the exact opposite of what is true.

 

 

25 minutes ago, Sauron said:

It also doesn't always directly correlate with well known application specific benchmarks.

CPUs are general purpose by design. They are designed to be able to do a very wide range of tasks, and they often do. Let's leave the IT realm and look at tools instead. Let's say the M1 was this tool:

 

and you were in charge of comparing it to this tool:

 

How would you do it?

I would test each individual tool, give them a rating from 1 to 10 and then take the average. That's what Geekbench does.

What you are saying right now is "no that's not a good way of comparing these tools with one another. What we should do is wait for someone to review specifically the bottle opener and then we should form our opinions based on that test".

Sure, the "testing all tools and taking the average" might not be a good test if you only care about the bottle opener, but if you are going to use every tool available (which you will on a CPU) then you need to test all of them to get the bigger picture of which tool is the best.

 

 

  

25 minutes ago, Sauron said:

Cinebench reflects performance in a specific application and is known to track pretty well with other intensive workloads. Because it's actually doing something and not just running a bunch of synthetic math that's supposed to simulate that something it has a real world meaning on what performance you can expect it to offer. Of course, just one benchmark would still not be enough to show you the whole picture but I'm inclined to trust it a lot more than geekbench.

I strongly recommend you do a bit of research before commenting on something. This post right here, demonstrates very clearly that you have 0 understanding or knowledge of Geekbench.

 

Geekbench is almost entirely built using real world programs. It is not a simulation of real world workloads. It is real world workloads. 

Wanna know how Geekbench tests PDF rendering? It opens a PDF using PDFium, which is the library used in many programs including Chrome.

Wanna know how it does its SQLite test? It loads a database into memory and uses SQLite (the worlds widely used database engine) to do queries against that database.

Wanna know how Geekbench does its compile test? It compiles a C source code file using Clang, one of the most widely used compilers in the world.

Wanna know how it does its HDR test? It does it by taking 4 SDR images and running it through the HDR algorithm described in this paper, which is how most HDR images are created.

The text compression test? It uses the LZMA SDK. It is also used in 7-Zip.

 

The list goes on.

Geekbench is literally a test suite created by ripping out components from popular, real world programs. It it exactly the same as how Cinebench is based on CINEMA 4D, except Geekbench is a much wider and more general purpose suite.

[Sauron]

2 minutes ago, LAwLz said:

Geekbench gets routinely blasted by people who doesn't have the technical understanding necessary to actually make the claims they make.

It's like saying "vaccines gets blasted on Facebook whenever they get brought up". Just because some loud and vocal minority who doesn't understand something says something is bad doesn't mean it is bad.

It's important to understand what benchmarks does to analyze the data they provide.

It's funny you'd say that considering we gave you plenty of justification for our distrust and you have yet to provide a single counterexample.

4 minutes ago, LAwLz said:

[Citation Needed] on it giving misleading results across architectural barriers.

https://www.forbes.com/sites/patrickmoorhead/2015/06/12/misunderstood-or-inappropriate-mobile-benchmarks-are-hurting-the-industry-and-consumers/

5 minutes ago, LAwLz said:

That's what a general purpose benchmarking suite is suppose to do. Test a wide range of things and aggregate them into a generalized score. That's pretty much the only way to generalize the performance of a general purpose CPU. It doesn't make sense to do it any other way if you want a general overview of the performance of a processor. You have to do a wide range of tests since a CPU is purposely designed to do a wide range of tasks.

Except that CPUs are only partially general purpose machines. If performance could be generalized why run more than one test in the first place? Most CPU architectures (and SoCs in particular) have special hardware sections designed for specific tasks that are faster than just running that task on the general purpose sections.

 

Hypothetically a CPU could be 10 times faster than the competition on, say, text compression due to dedicated hardware but significantly slower on everything else, and an aggregate benchmark would give you a result that is misleading in EVERYTHING: it would penalize that CPU regarding text compression and it would penalize other cpus in everything else. It's just a bad idea to take results from radically different benchmarks and just average them out to get a leaderboard. It's the same problem websites like userbenchmarks have; it's just a bad way of measuring these things.

21 minutes ago, LAwLz said:

[Citation Needed] on Apple pushing their ARM CPUs to the absolute maximum thermal limit.

That might be true for their x86 products, but I have seen no indication of Apple pushing their ARM CPU cores to the max in terms of thermals.

Oh ok, I guess we can just ignore their pattern of behavior with almost every single laptop they've ever released because it's ARM this time? Note that I just said they have a reputation of doing this, not that they DEFINITELY DID THIS here. Which is also why more and longer benchmarks would be pretty useful before coming to a conclusion.

23 minutes ago, LAwLz said:

Also, like I said earlier, Apple's ARM chips does not have any boosting capability. The fact that the tests are short is a DRAWBACK when testing these chips against x86 processors.

Just because it won't boost beyond the stated base frequency doesn't mean it's not going to throttle at high temperatures... those are separate things. I find it unlikely that a chip in a laptop without fans wouldn't have thermal protections. This time you're the one who'll have to provide a citation for x86 chips being favored...

27 minutes ago, LAwLz said:

I would test each individual tool, give them a rating from 1 to 10 and then take the average. That's what Geekbench does.

And that would be a pretty bad way of measuring it for reasons I stated above. Why take the average? Furthermore the way you would "test" and weigh each individual tool would skew the result in favor of one or the other.

29 minutes ago, LAwLz said:

What you are saying right now is "no that's not a good way of comparing these tools with one another. What we should do is wait for someone to review specifically the bottle opener and then we should form our opinions based on that test".

Do you know what correlation is? It turns out that if you measure certain things you get results that are meaningful across a variety of different tasks. In your analogy this might be akin to measuring the quality of the materials used to give an estimate of how well they're all going to behave while still admitting the possibility that special adjustments might make any individual tool better or worse than predicted. On the other hand measuring each tool in unrelated tasks might give you that missing ultraspecific information but not the general quality of the tool for uses that weren't specifically accounted for.

[IceCaveMan]

1 hour ago, LAwLz said:

That is true. 

It's worth noting that Apple's ARM processors does not have a boosting mechanic, so the point Linus brings up is an advantage to x86 and a disadvantage to ARM.

True. Linus has been playing a bit like Bryson DeChambeau at the Masters this week. I keep losing golf balls with my shots? I'll just hit it harder!

 

He's not playing from the fairway and is well below the cut line. In truth it's a bit embarrassing for him...his emotion is driving him to break his own golf clubs. Painful to watch such a meltdown.

[Orange1]

12 hours ago, LAwLz said:

to make a list of all the idiotic and stupid comments

that Linus has made. People slip up, no big deal, move on.

 

I dont see a problem with either but It'd be a waste of time.

[Video Beagle]

On 11/14/2020 at 2:11 AM, Orange1 said:

 

I think Linus is just playing a role, and Linus understands well that drama sells.

Yet Marques' video that came out at the same time covered the same issues, without stoking drama, without a click bait headline, has over twice as much views.  Drama sells to a point..maybe profesionalism sells more.

[Video Beagle]

19 hours ago, LAwLz said:

If building PCs is the IT world's burger flipper at McDonalds, then Linus is the IT world's Ronald McDonald.

Ouch.  Funny, but ouch.

I will say in defense of Linus, who I will criticize a lot, he is very good at presenting and educating in an entertaining format, and that's a very valuable skill. Having all the knowledge in the world is pointless if you can't communicate it in a way that people will listen to. 

Still funny, though, @LAwLz

[Spindel]

OMG can the internet please stop jumping on Linus for his opinions on this.

 

His first video I still think was good and fairly well balanced, except for the title (he has explained the click bait titles on many occations).

 

I had to watch thhe WAN show after the shit storm it stired up, and I still think Linus opinions are OK.

 

I may not share his viewpoint all the time but I still think he does a good job in reporting on this. Way betters than most Apple haters on this forum.

[LAwLz]

6 hours ago, Sauron said:

It's funny you'd say that considering we gave you plenty of justification for our distrust and you have yet to provide a single counterexample.

No you haven't. You have said things that are completely wrong and then don't acknowledge that you are wrong. You have pretty much 0 understanding of Geekbench and you refuse to listen. I can't provide counterexamples because the things you are saying are wrong. It feels like talking to an anti-vaxxer that are saying vaccines are causing autism and now it's somehow up to me to prove that they aren't? 

 

6 hours ago, Sauron said:

https://www.forbes.com/sites/patrickmoorhead/2015/06/12/misunderstood-or-inappropriate-mobile-benchmarks-are-hurting-the-industry-and-consumers/

Did you even read the article you are linking? It literally says Geekbench is a "bad" benchmark because it only tests CPU and memory, not things like the GPU. It doesn't say Geekbench shouldn't be used to compare CPUs of different architectures. In fact, the article says THE EXACT OPPOSITE. For crying out loud do a tiny bit of research before posting, please. I beg you. At the very least read your own damn links.

Quote

Geekbench

 

Geekbench is a cross-platform benchmark that started out on MacOS and iOS, popularized by its ability to both run on iOS and Android and provide some feedback about certain aspects of a CPU’s architecture and theoretical capabilities. It is currently on version 3 and now supports testing on Windows, Mac, Linux, Android and iOS. Geekbench isn’t as bad of an offender as AnTuTu when it comes to being a misleading or misunderstood benchmark, but it does only test two components of a smartphone, the CPU and memory, and doesn’t do so in any real world scenarios. This leaves out really important components like the GPU, which is fast-becoming a compute workhorse in a heterogeneous compute environment or storage.

 

The tests it runs are CPU integer and floating point calculations in both single core and multi-core modes as well as memory single core and multi-core. As a result, the benchmark may provide some insights as to the architectural comparison of the CPU in the system, but in the case of two different smartphones with the same SoC this benchmark provides limited to no value. This benchmark should only really be used to compare CPUs from different operating systems and platforms and how they stack up against each other, not a benchmark for comparing phones, especially not for a review.

 

Geekbench certainly has its place in benchmarking, but it doesn’t particularly make sense to be including it the way that smartphone reviewers have been doing in their reviews. As a result, this becomes an inappropriate mobile benchmark because of the way it gets utilized by the press. Geekbench should be commended for their transparency of how and what they’re testing exactly, but it doesn’t change the fact that it’s being misused in mobile testing.

The article you just linked agrees with me and says that you are wrong. 

This is why I am so irate in this thread as well as the other M1 thread. The people arguing against me have no idea what they are talking about and doesn't even read their own sources. 

 

 

7 hours ago, Sauron said:

Except that CPUs are only partially general purpose machines.

No they aren't.

 

7 hours ago, Sauron said:

If performance could be generalized why run more than one test in the first place?

Do you not understand what general purpose means? General purpose means it can do multiple tasks. That's why you need to run multiple tests to get a general understanding of how it performs.

 

 

7 hours ago, Sauron said:

Most CPU architectures (and SoCs in particular) have special hardware sections designed for specific tasks that are faster than just running that task on the general purpose sections.

Yes but how is that relevant when we are evaluating the general purpose parts (the CPU)?

I am not sure if it was you but someone arguing with me earlier said that Geekbench was bad because it didn't use the general purpose execution units on the M1, and now you are saying it is bad because it isn't using the general purpose execution units?

The general purpose part of the M1 are fantastic. On par with Zen3 from the looks of it. But in addition to having great general purpose performance, it also seems to have better and wider support when it comes to specific functions in fixed hardware.

No matter how you twist and turn your arguments, the M1 seems to be a fantastic chip.

 

 

7 hours ago, Sauron said:

Hypothetically a CPU could be 10 times faster than the competition on, say, text compression due to dedicated hardware but significantly slower on everything else, and an aggregate benchmark would give you a result that is misleading in EVERYTHING: it would penalize that CPU regarding text compression and it would penalize other cpus in everything else. It's just a bad idea to take results from radically different benchmarks and just average them out to get a leaderboard. It's the same problem websites like userbenchmarks have; it's just a bad way of measuring these things.

How is this relevant to Geekbench or let's say SPEC?

It isn't running on dedicated hardware. The AES test might be doing that, but even if it is, both Intel and AMD also have dedicated hardware for that.

 

 

7 hours ago, Sauron said:

Oh ok, I guess we can just ignore their pattern of behavior with almost every single laptop they've ever released because it's ARM this time? Note that I just said they have a reputation of doing this, not that they DEFINITELY DID THIS here. Which is also why more and longer benchmarks would be pretty useful before coming to a conclusion.

We already have longer benchmarks of the iPhone, which uses a chip very similar to the M1.

Why do you think the M1 in the Macbook will thermally throttle when the A14 in the iPhone doesn't?

It's entirely possible that the M1 might thermally throttle, but I think the risk of it doing that are extremely small to the point where it doesn't make sense to assume it will.

 

On 11/14/2020 at 5:23 PM, Sauron said:

Just because it won't boost beyond the stated base frequency doesn't mean it's not going to throttle at high temperatures... those are separate things. I find it unlikely that a chip in a laptop without fans wouldn't have thermal protections.

Of course. I never said the opposite (that boosting and thermal throttling being separate things). Not sure why you bring it up.

Also, two of the three devices announced do have active cooling. Only one of the laptops do not, and that laptop will most likely have the entire chassi as a heatsink. If the iPhone can handle the heat in its tiny chassi, I don't think there is a big risk of the Macbook not being able to handle it.

The active cooling in the Macbook Pro and the iMac will probably help though, but I don't think the regular M1 Macbook will have like massive thermal issues either.

 

 

On 11/14/2020 at 5:23 PM, Sauron said:

This time you're the one who'll have to provide a citation for x86 chips being favored...

You said Geekbench was bad because the tests were short and that would allow the tests to finish during boost periods.

ARM chips do not have boosting capabilities, x86 do. That means that, in your own words, boosting the entire test leads to unrealistically high benchmark results. That only applies to x86 in Geekbench since ARM do not have that feature. That means that the whole "the test is short so chips can boost and get higher scores than they should" only applies to x86, not ARM. Thus, x86 has at least one benefit in Geekbench.

Your argument makes no sense.

 

On 11/14/2020 at 5:23 PM, Sauron said:

Do you know what correlation is? It turns out that if you measure certain things you get results that are meaningful across a variety of different tasks. In your analogy this might be akin to measuring the quality of the materials used to give an estimate of how well they're all going to behave while still admitting the possibility that special adjustments might make any individual tool better or worse than predicted. On the other hand measuring each tool in unrelated tasks might give you that missing ultraspecific information but not the general quality of the tool for uses that weren't specifically accounted for.

I have no idea what you mean or what your point is. Can you elaborate?

Please remember that Geekbench do test a wide range of real world programs, and also breaks down each separate benchmark into an individual score which can be used. If you don't like the whole "let's aggregate everything into one score", you can just look at the individual scores from the individual programs.

[LAwLz]

10 hours ago, Spindel said:

OMG can the internet please stop jumping on Linus for his opinions on this.

 

His first video I still think was good and fairly well balanced, except for the title (he has explained the click bait titles on many occations).

 

I had to watch thhe WAN show after the shit storm it stired up, and I still think Linus opinions are OK.

 

I may not share his viewpoint all the time but I still think he does a good job in reporting on this. Way betters than most Apple haters on this forum.

I haven't looked at any of Linus' videos yet. I am currently looking at the WAN show (18 minutes in) and it is very clear that Linus doesn't know what he is talking about. What makes this even worse is that he has this smug attitude of "I got nothing wrong", and then goes on to shit his pants when talking about video encoding on the M1 vs Intel Macs.

 

Apple can't build the fixed function hardware to handle a bunch of different codecs and settings? Well good thing for Apple that it seems like the M1 will handle doing video encoding and decoding in software better than ~4 core Intel chips, especially at the thermal restrictions typical for laptops and other portable devices.

 

 

Linus and several people on this forum seems to be under the impression that the M1 has really weak CPU wise, but makes up for it with fixed function hardware for specific tasks. That is not the case. The M1 is REALLY powerful CPU wise, in addition to having some fixed function hardware (but not that much more than Intel's CPUs have).

 

 

Also, Linus saying these new Macs are just "iPads with a higher thermal target" is either him trying to be funny or mocking Apple. The real difference between the iPad and these new Macs is the OS they run. Just because two systems share the same CPU architecture (remember, the M1 and A14 isn't the same chip or even league of performance) does not mean they are "the same". Likewise, my smartphone isn't a super computer just because it shares the same instruction set as the Fugaku super computer. Nor is my GTX 1060 "essentially a 1080 Ti" because it shares the same architecture.

As Luke said, a higher thermal target can change a lot. But a completely different OS and roughly twice the amount of performance can change things too, and that changes things far more.

 

Edit:

Oh God Linus... The more I listen to the WAN show the more things Linus gets wrong. As soon as I had written that part about Luke talking about thermal targets I unpause the video and what do I hear? Linus talking about boosting certain cores, WHICH ISN'T A THING ON ARM.

Linus should shut his ignorant mouth and do a bit of research before spewing garbage to his hundreds of thousands of viewers.

 

 

Edit 2:

At 29:29 Linus says:

Quote

You know why we don't use Geenbench in our reviews? Like, unless we are working with a Mac or trying to do a cross-platform comparison? Do you have any idea why we don't do that? Because it's not a very good benchmark. It's just nor particularly meaningful.

Quote

1) It doesn't run for very long so it doesn't tell you anything about thermal throttling.

This is a benefit for x86 which can turbo during these short segments, a luxury ARM doesn't have.

The benchmarks being short is a benefit to x86 and a drawback to ARM. Not the other way around as Linus seems to imply here. It's not a secret that ARM processors typically produce less heat that x86 chips. So I am not sure why Linus brings this up as a reason for why Geekbench is a bad benchmark when comparing the M1 against an AMD or Intel CPU.

Quote

2) It's very synthetic. You know it's not something we would consider real world.

Things like compiling with Clang, opening PDF documents with Chrome, doing shortest path calculations with dijkstra's algorithm and compressing text files with 7-Zip is apparently not "real world" enough according to Linus... The same Linus that uses Cinebench in his review.

For crying out loud Linus even used SPEC in his Ryzen 5000 review, which is very similar to Geekbench.

 

 

I agree with Linus on some things though. For example I don't think the M1 Macbook and Mac Mini will age that well. A lot of first gen products don't. I wouldn't be surprised if the M2 is a lot better. For everything I agree with Linus on regarding the M1 though, he seems to get like 2-3 things wrong.

[Orange1]

10 hours ago, Spindel said:

OMG can the internet please stop jumping on Linus for his opinions on this.

Do you realize that will never happen!

 

It's like when famous people bitch and moan about being famous and not having privacy.

 

 

 

 

[Dracarris]

19 minutes ago, Orange1 said:

Do you realize that will never happen!

 

It's like when famous people bitch and moan about being famous and not having privacy.

As a very well-known tech-"influencer"/Youtuber/whatever, why should that happen? You represent a resource people turn to when they want to get informed about new tech. You obviously cannot put out a random opinion without factual backup when being such a person. You have a certain responsibility that LS certainly usually lives up to, e.g., see the PS5 SSD apology-video. The reactions on Twitter from Sony and Epic executives also reinforces the fact that he is too important and well known to spit out BS to millions of people.

[RedRound2]

17 hours ago, Spindel said:

OMG can the internet please stop jumping on Linus for his opinions on this.

 

His first video I still think was good and fairly well balanced, except for the title (he has explained the click bait titles on many occations).

 

I had to watch thhe WAN show after the shit storm it stired up, and I still think Linus opinions are OK.

 

I may not share his viewpoint all the time but I still think he does a good job in reporting on this. Way betters than most Apple haters on this forum.

People have said this but I'll reiterate again and tag @LinusTechif he ever bothers to see this and the others who replied to you

 

When you are tech influencer, with millions of viewers relying on you for information, you have a duty to make sure what you say is correct and accurate. 

 

And to be clear, it's only his opinion if there's something subjective about the matter. But there isn't. It's an objective measurement that he jumped the gun and portrayed it as an iPad

 

I get the skepticism around performance claims. I get it. Thats alright. But to straight up calling it an iPad chip - to regular people means it's just an iPad which is far far from true (like really does he not understand it's akin to calling a full grade desktop tower's i7 processor equivalent to a ultrabook i7??!. And in Apple's case they even have different names, A series and M series!)

 

MKBHD made a similar skeptical video about the M1 and I'm perfectly fine with it. Linus just went full on decided what the M1 was going to be with the title and his very smug "yeah right lol" attitude

 

Regarding to the lack of detailed performance charts, when has anybody taken a manufacturers claims to actual value. They're more often than not misleading in some way or form. Plus Apple has never really compared directly to competition like that, at least in hardware. They usually compare the new chip to themselves - which in this first gen they obviously don't have anything to compare to.

 

Also, Apple's customers don't tend to care about of any of these things nor are they technically inclined like us to understand. Apple actually did give more details than usual with their M1 picture below, and from my research the 8 bit ultra wide execution and all parts of SoC having direct access to DRAM (hence unified memory) is one of the few things that enables them to have a huge performance uplift

 

Linus has a tendency to make assumptions and have a pre notional verdict on things he that seems unexpected. He has done it so many times before, with recent example being PS5 SSD - where a bit of research or an attempt to understand the claims would've gone long way from preventing one to say such unsubstantiated opinions and dissing that he usually tends  to give

 

[Orange1]

10 minutes ago, RedRound2 said:

He has done it so many times before

 

10 minutes ago, RedRound2 said:

where a bit of research

 

10 minutes ago, RedRound2 said:

or an attempt to understand the claims would've gone long way from preventing one to say such unsubstantiated opinions and dissing that he usually tends  to give

 

 

Yeah, well the push to spit out vids fast with no time for research. Get it while its hot!

 

Yeah, I wouldnt take the manufacturers word on graphs. Have to do several rounds of "research" from several different yt'ers to see whats true and was marketing lingo.

[Video Beagle]

22 hours ago, Spindel said:

His first video I still think was good and fairly well balanced, except for the title (he has explained the click bait titles on many occations).

Actually, and again, speaking as an Apple user since 1980, I thought the video was generaly fine. But, as you said, the TITLE is a problem. And, as I said above, MKBHD did a video with the same info, no click bait title, and when I last looked had over 3 million views to LTT's 1.2, so I don't think the "click bait" excuse holds water as much, especially when you're in the top..what 5?3? of tech youtube channels.

 

MKBHD and Snazzy Labs (wearing a LTT hoodie, fwiw) did videos at the same time LTT did, but mentioned the footnotes (and I think SL DID have the model of computers that were being compared, not just the processor as was claimed on WAN). They didn't have the clickbait title, but expressed the same doubts and complaints about the clear marketing material presented.

 

The thing is, and others have pointed to this, Linus spent 40 minutes on WAN trying to justify his postiion (with luke providing his usual sidekick support). Which it wasn't his position that's the problem...it's his ..atitude....He says not to fan boy a company with one breath, and then anti-fan boys apple with the next. He uses their products that he likes and thinks are good (such as the air pods) but tries to downplay it (like a constant "joke" he says about the apple watch that he wears it ironically or something because android watches are bad?  I can't remember the term he uses)...HE could just say "I like it.".

 

He rightfully has pride in his company and the postiion that they've earned in the tech press/influencer space, but he also seems to really resent that Apple just doesn't care. Now, I don't know what influencers Apple does deal with (MKBHD? iJustine I'm pretty sure)....but Apple is more than a tech company...as it's detractors like to yell out, it's also about image.  And just maybe putting out multiple videos in a week with click bait titles about then, while him and his staff mockingly call the top gay CEO in the world the same name used by the world's biggest homophobe just doesn't project the image that Apple looks for.

[Spindel]

7 minutes ago, Video Beagle said:

Actually, and again, speaking as an Apple user since 1980, I thought the video was generaly fine. But, as you said, the TITLE is a problem. And, as I said above, MKBHD did a video with the same info, no click bait title, and when I last looked had over 3 million views to LTT's 1.2, so I don't think the "click bait" excuse holds water as much, especially when you're in the top..what 5?3? of tech youtube channels.

 

MKBHD and Snazzy Labs (wearing a LTT hoodie, fwiw) did videos at the same time LTT did, but mentioned the footnotes (and I think SL DID have the model of computers that were being compared, not just the processor as was claimed on WAN). They didn't have the clickbait title, but expressed the same doubts and complaints about the clear marketing material presented.

 

The thing is, and others have pointed to this, Linus spent 40 minutes on WAN trying to justify his postiion (with luke providing his usual sidekick support). Which it wasn't his position that's the problem...it's his ..atitude....He says not to fan boy a company with one breath, and then anti-fan boys apple with the next. He uses their products that he likes and thinks are good (such as the air pods) but tries to downplay it (like a constant "joke" he says about the apple watch that he wears it ironically or something because android watches are bad?  I can't remember the term he uses)...HE could just say "I like it.".

 

He rightfully has pride in his company and the postiion that they've earned in the tech press/influencer space, but he also seems to really resent that Apple just doesn't care. Now, I don't know what influencers Apple does deal with (MKBHD? iJustine I'm pretty sure)....but Apple is more than a tech company...as it's detractors like to yell out, it's also about image.  And just maybe putting out multiple videos in a week with click bait titles about then, while him and his staff mockingly call the top gay CEO in the world the same name used by the world's biggest homophobe just doesn't project the image that Apple looks for.

I just have to put in that a thing I like with iJustine is that she always focus on the positive. She does not focus on negativity. (Also I love how she tried to poach Teran during their editing showdown  )