m1 Apple M1 = the rest of us are living in the stone age!?

[Spindel]

6 hours ago, Donut417 said:

Well firstly its highly unlikely Apple will sell its chips to third parties. So all we can hope for is that this pushes Qualcomm, Samsung, Nvidia, etc to put more money in to R&D and make a chip that is just as good or better. The other side of the coin is Microsoft. One of the things that has helped make Apple's transition successful is Rossetta 2, while Microsoft can emulate X86 apps on ARM its not as good as what Apple has achieved, from what I have read. Microsoft needs to step up its game. 

 

On a second note, we dont know what innovations AMD and Intel have for us in the future. That being said, I think that if Apple can be successful and other ARM chip makers see this, it could have potential to directly compete with Intel and AMD. Assuming Microsoft can do better. To me where ARM would shine in the PC space is Laptops. Due to its power efficiency and the fact companies like Qualcomm could potentially build 4G and 5G radios in to the device. 

 

I think once we get to see benchmarks of Apples desktop line up its going to give us an idea of what we can expect in the future as well. As some have mentioned, ARM could very well be the future. But that day isn't here yet and I think it will be years before we really see a push. 

About Rosetta:

It actually is really fantastic on the M1. Of all the apps I have run I wouldn't even know they where compiled for Intel unless I look in the Activity monitor. And in my case any performance loss is not noticeable in daily use. Of course YMMV depending on what apps you use (and of course some apps do not work with rosetta, but non of the ones I've needed have failed). 

 

The thing I do notice is that once an update of an app to native AS is made is that the RAM usage for that app drops significantly. In the instances where I have looked for it it has been 30-50 % less RAM hogging.  

[Video Beagle]

16 hours ago, GDRRiley said:

M1 being on 5nm is whats keeping it in the race. if it was on 7nm or someone other than TSMCs node zen3 would be beating it. Zen would easily be ahead by around 10% and m1 would draw more power.
it would have been so good for capacity if Zen3 APUs were on 5nm right now.

 

Your argument is "if they were different things than they actually are, the results would be different"?

 

[A51UK]

2 hours ago, gal-m said:

I'm really looking forward to that day @Donut417!

Really interested in Microsofts response if Apple Desktop Silicon turns out to actually be good, which I believe it will.

Microsoft do not need to response as it is software company not hardware company. They got Windows 10 on ARM now, may make on sale for public if ARM CPU become more on desktop. I think Microsoft would or need to care very little about  Apple Desktop Silicon as it make basic no different to them. They can put Windows on basic on any CPU if need to or want to.  Apple Desktop Silicon is not a real bit deal. 

[Nena Trinity]

Can it run Fortnite? War Thunder? Rainbow Six Siege?

[Donut417]

1 hour ago, A51UK said:

They can put Windows on basic on any CPU if need to or want to.  Apple Desktop Silicon is not a real bit deal. 

Yes and no. Your forgetting that Businesses dictate what Microsoft can do. Thats what happens when you have the most market share. Microsoft is used by big business and the government. So that software library they have accumulated would need to work. Not all those devs who made that software still exist OR are going to be willing to recode that software. Yeah Microsoft has Windows on ARM, but its not any where close to what Apple has released. Microsofts emulation of X86 is not as good as what Apple has. 

 

The reason Apple silicone is a really big deal is that in some work loads it does better than X86, but its only sips power. As you said, they are making their own ARM chip. But unlike Apple who has 10+ years doing CPU design Microsoft doesn't have that level of experience. So I doubt their first CPU is going to be this impressive. 

 

Microsoft will be the primary reason if ARM succeeds on PC. Apple has too little marketshare to really matter here. Linux pretty much runs on anything but is mostly server based. 

[Donut417]

7 minutes ago, Nena Trinity said:

Can it run Fortnite?

No because Apple hates Epic and if that hate is anything like they have for Nvidia, they will make sure Fortnite is never supported. 

[LAwLz]

19 hours ago, GDRRiley said:

it is worse and you just prove it right there.
M1 being on 5nm is whats keeping it in the race. if it was on 7nm or someone other than TSMCs node zen3 would be beating it. Zen would easily be ahead by around 10% and m1 would draw more power.
it would have been so good for capacity if Zen3 APUs were on 5nm right now.

You keep moving the goalpost.

Also, N5 isn't what is keeping the M1 in the race. As we saw in the tests I linked above, AMD's best chip, the 5980HS, is not able to keep up with the M1 in terms of performance or power consumption. Even if we were to take TSMC's best case numbers and reduce the power consumption by 30%, or increase performance by 15%, AMD's processor still would use more power than Apple's, and only be able to match it.

It's very important that you understand this. The M1 uses less power, and is faster than AMD's best chip, and the gap is big enough than a node shrink would not really change anything.

 

AMD, even with a node shrink, would have to choose. Do we want the same performance as Apple? In that case we will use more power. Do we want the same power consumption as Apple? Then we don't get the performance.

[A51UK]

3 minutes ago, LAwLz said:

You keep moving the goalpost.

Also, N5 isn't what is keeping the M1 in the race. As we saw in the tests I linked above, AMD's best chip, the 5980HS, is not able to keep up with the M1 in terms of performance or power consumption. Even if we were to take TSMC's best case numbers and reduce the power consumption by 30%, or increase performance by 15%, AMD's processor still would use more power than Apple's, and only be able to match it.

It's very important that you understand this. The M1 uses less power, and is faster than AMD's best chip, and the gap is big enough than a node shrink would not really change anything.

 

AMD, even with a node shrink, would have to choose. Do we want the same performance as Apple? In that case we will use more power. Do we want the same power consumption as Apple? Then we don't get the performance.

How do you work it out?  I think AMD could get more or same performance at the same node level as the M1, The AMD choose if do they want more higher performance at the cost of power consumption or just a bit high performance at the same power consumption as Apple. ARM does get better power consumption but at lower performance in the past, cannot see it going be any different now.  AMD and Intel over the years most pick performance over power consumption. 

 

 

[Donut417]

7 minutes ago, A51UK said:

cannot see it going be any different now. 

Before now you didn’t have a company worth over a trillion dollars pushing it. Also I don’t think any of the ARM chip makers expected their chips to end up in PC/MAC. They were content just supplying the phone market. But Apple was done with Intel and Intel’s bullshit. This is why performance will get better, because Apple needs it to be. 
 

The question that’s not being asked is how does this effect other ARM chip makers? Because Apple isn’t competing with Intel or AMD. Apple is a closed eco system. It’s going to be Qualcomm, Nvidia and or etc who compete with Team Red and Blue. 

[NunoLava1998]

30 minutes ago, LAwLz said:

You keep moving the goalpost.

Also, N5 isn't what is keeping the M1 in the race. As we saw in the tests I linked above, AMD's best chip, the 5980HS, is not able to keep up with the M1 in terms of performance or power consumption. Even if we were to take TSMC's best case numbers and reduce the power consumption by 30%, or increase performance by 15%, AMD's processor still would use more power than Apple's, and only be able to match it.

It's very important that you understand this. The M1 uses less power, and is faster than AMD's best chip, and the gap is big enough than a node shrink would not really change anything.

 

AMD, even with a node shrink, would have to choose. Do we want the same performance as Apple? In that case we will use more power. Do we want the same power consumption as Apple? Then we don't get the performance.

It's not faster than the 5980HS; it's the same in terms of single core, but it falls behind by about 40% in multicore (atleast on cinebench r23).

However, it's still magnitudes better than the 5980HS or any other mobile x86 CPU in terms of performance per watt; a shrink to 5nm on x86 still wouldn't get close 

[LAwLz]

18 hours ago, leadeater said:

I've said it before and I'll say it again, Geekbench is the most garbage "benchmark suite" that has always and continues to extremely badly translate across to actual real application performance. If you use the actual few application tests to compare them there is very little between the two for single thread performance.

You keep saying it but never provide any evidence or reason why you think that.

Could it be that you just dislike it because it doesn't show you the results you want it to show? 

 

 

18 hours ago, leadeater said:

Is this a problem with Mac OS platform itself, sure and that's not your fault, it is however yours to actually take any meaningful credit of Geekbench.

I have also linked other benchmarks if you don't like Geekbench.

 

18 hours ago, leadeater said:

And no I'm not discrediting it because the M1 is a decent bit faster in it, I'm discrediting it because it's had this problem well before M1 was a thing. Was crap back then and it's still crap now.

Such as? And why do you ignore the other benchmarks?

 

18 hours ago, leadeater said:

And can we please stop acting like single thread performance is the only thing the matters, zero people buying a 5980HS or any other mobile or non mobile high core count CPU is not buying it for single thread performance.

What's funny is that when I said zen3 wasn't that amazing because you could get higher core count zen2 products for the same price you dismissed that and made the argument that single core was more important.

Now that the tables have turned and AMD aren't on top of single core you suddenly say we should focus on multi-core performance? 

Remember a couple of months back when you kept saying "look at zen 2 single core performance and apply 20% which we will get to zen 3 and it isn't that impressive"? 

I find it interesting that you kept talking about single core performance before but now when Zen 3 mobile doesn't live up to your expectations (like I said it wouldn't) you don't care about it anymore?

 

 

18 hours ago, leadeater said:

Because that's just the thing, if Intel or AMD wanted to greatly widen their cores to achieve the same thing they could, they don't because it doesn't lend very well when considering high core count scaling and some of the draw backs of x86 itself.

You make it sound like Intel and AMD choose to not go "as wide" as Apple because they don't care about single threaded performance, but could if they wanted to.

Firestorm aren't impressive just because they are wide. I mean, if width was the only thing that mattered then Samsung's Mongoose cores would be great too, but they aren't.

Besides, making a wide decoder is far easier with the ARM ISA than the x86 ISA since all instructions are the same length (8 bits).

You can't really just point to "it's the width that makes it" when CPU design between x86 and ARM are like apples and oranges.

 

19 hours ago, leadeater said:

And nobody has said otherwise. That's the benefit both Intel and AMD have, the fact that they have dynamic clocks and even more so for AMD the power management on Zen allows it to cater for a very broad range of workloads giving the best performance it can. The fact that as core count and core utilization goes up the clocks goes down doesn't matter as this reduction is non linear and as we have already seen and still see multithread performance is higher.

The problem is that some people act like AMD can have it both ways, which they can't.

For example some people have said Zen 3 can go as low in power consumption as Firestorm cores, but then don't mention (maybe deliberately "forget" to mention it) that the performance of zen 3 cores at that frequency falls behind the Firestorm cores.

[LAwLz]

19 hours ago, leadeater said:

False, I gave you the evidence, seems you just ignored it because you simply didn't want to believe it. You can go back to Anandtech and look at the per core power scaling across clocks again if you like, if you still don't believe it then your denying reality. They even have the IOD power nicely separated for you.

 

Then you can cross reference those figures with APUs and mobile parts that use a monolithic die and contain less IO and don't have to maintain IF link between chiplets and see that in this application power is indeed much lower.

No you didn't...

You posted a picture of isolated core measurements and then made the assumption that because going from 15 active cores to 16 active cores only increases power consumption by let's say 5 watts, a zen 3 core only consumes 5-6 watts and therefore you attributed the other wattage to IO. THAT IS AN ASSUMPTION.

Do you have measurements of how how power is consumed by the IO die? Not "I believe it uses this much because other numbers shows this and I think the remaining power is used by the IO because I personally can't think of anything else that could use it", but actual hard numbers for it.

 

How many times do we have to go over this?

 

0 cores loaded - 11 watts of power

1 core loaded - 28 watts of power (17 watts more)

2 cores loaded - 41 watts of power (13 watts more) - The reason why going from 1 to 2 cores is a smaller number than 0 to 1 is because some shared resources are activated when core 1 is loaded, and thus they do not have to be activated for the core 2, and the chiplet is already being supplied appropriate voltage and power.

3 cores loaded - 56 watts of power (15 watts more)

Etc etc

 

You keep looking at numbers like "if 15 cores are already active then dropping the frequency on all 15 cores by 75MHz and activating the 16th core only results in the 16th core using 5 watts of power and therefore a zen 3 core only uses 5 watts of power" as some kind of evidence for something.

 

Facts are facts.

If you want to have a zen 3 core performs as well as a Firestorm core, then the core itself, not including IO or anything like that, will use about 15 watts of power. It's only when you drop the frequency way down that you get results like a Zen 3 core using 6 watts of power, and even in those scenarios, we are only looking at the core itself, when it already has things like the decode engine (which in x86 is fairly big and complex) and the chiplet lit up.

 

If you want to say zen 3 only uses 5-6 watts of power per core, then you are not allowed to say it performs as well as a Firestorm core, because in the 6 watt config it doesn't, not even close.

If you want to say Zen 3 performs as well as a Firestorm core then you aren't allowed to say it doesn't use more power because it does. A single zen 3 core at that performance level uses about 15 watts of power for the individual core alone, while the entire Mac Mini including fans, memory and everything uses about 22 watts for 4 big cores and 4 small cores.

 

And stop saying the IO die is the reason why AMD chips use more power than the M1. We don't know that. You might think it is the reason why, but I have so far seen literally 0 evidence to support that theory. I have no idea how you came up with that idea even.

The numbers I have seen circulating puts the IO die power consumptions at around 10 watts. Don't get me wrong, that is a lot, but I seriously doubt the very impressive efficiency if the M1 will suddenly go out the window as soon as they add more IO like some people think.

[NunoLava1998]

One thing I'd like to mention is that I bought the entry level M1 MacBook Air today (7-core GPU) and did some benchmarks on it; with a lot of other tabs open and on Google Chrome, I got around 61,000 on the Octane 2.0 benchmark:

And 640ms on Kraken 1.1:

These are some really impressive results; on the same browser, my Ryzen 5 1500X system got about 38,000 on Octane and 950ms on Kraken 1.1. Apparently they're on par with the i9-10900K(!).

 

In terms of GPU, the only benchmark I could really find for that was Geekbench, and it gave results pretty similar to a GTX 1050. Pretty good, and enough for some gaming if you really want to.

[LAwLz]

1 hour ago, A51UK said:

How do you work it out?  I think AMD could get more or same performance at the same node level as the M1, The AMD choose if do they want more higher performance at the cost of power consumption or just a bit high performance at the same power consumption as Apple. ARM does get better power consumption but at lower performance in the past, cannot see it going be any different now.  AMD and Intel over the years most pick performance over power consumption. 

I don't really understand what you mean. Can you please rephrase your question and statement?

Why do you think AMD would be able to match Apple's power consumption while providing better performance if they were on the same node?

 

 

1 hour ago, A51UK said:

ARM does get better power consumption but at lower performance in the past, cannot see it going be any different now.

But that hasn't been true even in the past. You are comparing the core designs from ARM (the company ARM, not the ARM ISA) to x86.

The CPUs Apple designs and uses are VERY different from the Cortex cores you see in SoCs Qualcomm puts out.

Making generalized statements like "ARM gets better power consumption but at lower performance" makes no sense because you're lumping very different CPU cores together.

 

 

55 minutes ago, NunoLava1998 said:

It's not faster than the 5980HS; it's the same in terms of single core, but it falls behind by about 40% in multicore (atleast on cinebench r23).

Where are you getting those numbers from? The Anandtech article I linked earlier shows this:

Geekbench single-core:

M1 - 1745

5980HS - 1506

Difference: M1 is ~16% faster.

 

Geekbench multi-core:

M1 - 7715

5980HS - 8391

Difference: M1 is ~8% slower

 

SPEC2006 geomean total:

M1 - 79.4

5980HS - 61.3

Difference: M1 is ~30% faster

 

 

I am having a really hard time finding decent benchmarks for the 5980HS.

[Spindel]

You plebs just have to accept that you are living in the storage, as the original intent of this thread was. 

 

I'm just waiting for you to argue that the sound is better in your cassette walkman than my digital music. 

[NunoLava1998]

1 minute ago, LAwLz said:

Where are you getting those numbers from? The Anandtech article I linked earlier shows this:

(...)

I am having a really hard time finding decent benchmarks for the 5980HS.

Cinebench R23 multi-core iirc is about 12,000 or so on the 5980HS, while on the M1 it's closer to 7,500.

I think the varying benchmark numbers are from just being on ARM; some benchmarks and applications take advantage of some of the ARM-only features (i.e. out of order execution), and some don't.

[LAwLz]

6 minutes ago, NunoLava1998 said:

Cinebench R23 multi-core iirc is about 12,000 or so on the 5980HS, while on the M1 it's closer to 7,500.

12,000 with the 5980HS doesn't sound too unrealistic but I'd still like to see some sources.

 

6 minutes ago, NunoLava1998 said:

I think the varying benchmark numbers are from just being on ARM; some benchmarks and applications take advantage of some of the ARM-only features (i.e. out of order execution), and some don't.

Just want to point out that out of order execution is by no means an ARM-only thing. That exists on x86 as well.

 

Another thing I would like to add is that I am not a big fan of Cinebench because it only tests one particular workload, while benchmarks like GeekBench and SPEC are benchmark suits testing a wide range of workloads.

If you make generalized statements like "processor X is better than processor Y", then you want to actually account for a wide range of programs, otherwise that statement might only be true for like 1 in 15 programs.

Running 1 program and making a generalized statement like "Processor X is the better overall processor, because of this one benchmark result" = Bad.

Running 15 programs and making a generalized statement like "Processor X is the better overall processor, because of these 15 benchmark results" = Good.

[NunoLava1998]

2 minutes ago, LAwLz said:

12,000 with the 5980HS doesn't sound too unrealistic but I'd still like to see some sources.

https://www.notebookcheck.net/AMD-Ryzen-9-5980HS-and-Ryzen-9-5900HS-vs-Intel-Core-i7-11370H-comparative-benchmarks-Tiger-Lake-H35-and-Cezanne-35-W-battle-it-out-for-single-core-supremacy.516875.0.html

[LAwLz]

13 minutes ago, NunoLava1998 said:

https://www.notebookcheck.net/AMD-Ryzen-9-5980HS-and-Ryzen-9-5900HS-vs-Intel-Core-i7-11370H-comparative-benchmarks-Tiger-Lake-H35-and-Cezanne-35-W-battle-it-out-for-single-core-supremacy.516875.0.html

Thanks for the link!

The Cinebench results seems really good for the 5980HS. But like I said earlier, I would like to see more benchmarks.

 

Apparently, NoteBookCheck has more benchmarks than they show in that article. If you look at their comparison tool you find stuff like WebXPRT, Kraken and 7-Zip as well.

 

Overall the M1 gets 88% of the performance of the 5980HS/5900HS in their tests. Although it is worth noting that some of the benchmarks they run seem to be running in Rosetta or a VM on the M1. For example Blender 2.79 doesn't support the M1 so they probably ran it in Rosetta. 7-zip doesn't exist for MacOS so my guess is that they used Keka, and I am not sure that is optimized for the M1 yet.

 

For a chip that is handicapped in a lot of tests, only has a quarter of the memory (they are testing the 8GB M1 vs the 32GB ROG Zephyrus), and uses about 70% less power, that's very impressive.

[leadeater]

5 hours ago, LAwLz said:

You posted a picture of isolated core measurements and then made the assumption that because going from 15 active cores to 16 active cores only increases power consumption by let's say 5 watts, a zen 3 core only consumes 5-6 watts and therefore you attributed the other wattage to IO. THAT IS AN ASSUMPTION.

No I did not, that is your assumption of what I said, go back and read what I said again and my explanations. We have direct measurements of the IOD power usage, we don't need to look at the per core usage but you can if you wish to but that has to do with the power efficiency of the cores themselves and the myth that Zen 2 or Zen 3 is not or cannot be power efficient and would somehow lose majority of it's performance when operating at more efficient clocks when it is which simply is not true, but obviously at lower clocks the cores will be less performant compared to the utmost peak.

 

I then gave you scientific journal articles from silicon engineers and designers who said exactly what I did, because that was my original sources for that information back when Zen 2 was launched and the IOD came to exist. However like I said back then those were not the exact sources I read because today I cannot find those specific ones but this was clearly explained back during Zen 2 launch and is part of the reason why the IOD is on GloFo 14nm and 12nm along with AMD having contractual obligation to use GloFo up until I think next year (not sure on that but it's expiring soon).

 

I've made zero assumptions on this statement so it comes back to you simply just not wanting to believe in the face of direct measurements readings and science plus experts saying the same thing. If you don't agree then you're just wrong, but if you wish to continue to be wrong you're welcome to but I invite you to be wrong in silence and not post lies like you did.

 

5 hours ago, LAwLz said:

And stop saying the IO die is the reason why AMD chips use more power than the M1.

No because when that was used by someone else because they only wanted to talk about desktop CPUs because there were no Zen 3 mobile parts that IS the reason why the idle power of the desktop parts is so high. You'd have to be blind not to see that.

 

You're totally lost and have no idea anymore of what people were saying at that time and have you own narrative stuck in you head. If you wish to discuss any such matters with me more then you're going to need to go back to those topics, start from the beginning and fix your complete misunderstanding of these conversations otherwise you'll continue to be clueless as to what my statements actually were and why they were made and to address what things. You've mixed up so many different things it's no wonder nothing makes sense anymore.

 

5 hours ago, LAwLz said:

If you want to have a zen 3 core performs as well as a Firestorm core, then the core itself, not including IO or anything like that, will use about 15 watts of power.

Correct, literally nobody has said otherwise. However what is your point with this? This is were you seem to get lost in all of this because nobody has said otherwise. What actually lead up to the conversation about Zen 3 per core power efficiency were claims, not by you but you joined in, from people that Ryzen Mobile is or was going to be many hundreds of percent higher power usage which is simply an illogical and incorrect statement to say which is why I addressed those specific claims with the per core clock scaling power.

 

In a mobile CPU with a lower TDP the clocks are going to be much lower, you can find those out by looking at reviews or any which way you like then go across to the Anandtech article find the closest MHz value in any one of those tables and look at the power the core is using. There you will see how much is actually being used not some ridiculously false hundreds of percent higher.

 

So when you load a Ryzen Mobile CPU with multiple threads it's power efficiency is much greater than if you were to load one core, that is the entire point of boosting, then with enough cores you have greater multithread performance than the M1 which is true, it was true even with Ryzen Mobile 4000. The only real difference between Ryzen Mobile 4000 and 5000 is single thread performance which increased by the amount indicated it would by AMD during the Zen 3 technical information releases. Now I will say I was expecting more of a multithread gain too but I can understand why, more work requires more power if no other changes i.e. node process.

 

Somehow at some point you've drift so far from this origin discussion I cannot make any sense of what you are saying anymore but it has nothing to do what I said or what was being discussed.

[leadeater]

6 hours ago, LAwLz said:

What's funny is that when I said zen3 wasn't that amazing because you could get higher core count zen2 products for the same price you dismissed that and made the argument that single core was more important.

Actually I said if you cared about gaming Ryzen 5000 was the better choice due to the much higher performance in actual games and made the most sense specifically for people who did not have a Ryzen CPU or a 1000 series. I did not say single core was more important, I said Ryzen 5000 made more sense to these buyers than Ryzen 3000 does today if you are buying multiple parts new that includes a new motherboard. If someone still values cost more than anything else Ryzen 3000 is still perfectly fine option.

 

Funny how you make up things to support only what you think and cast aside the reasons clearly stated to you. My advice, stop talking to me about Zen 3 and M1 until you can correct this issue. Frankly I'm finding it pathetic and it'll only lead to me just ignoring you other than to discredit any claims where you mention  my name and I'll be doing it in a much shorter fashion with a lot less effort.

 

Spoiler

  

6 hours ago, LAwLz said:

You can't really just point to "it's the width that makes it" when CPU design between x86 and ARM are like apples and oranges.

 

On 1/28/2021 at 8:03 AM, leadeater said:

some of the draw backs of x86 itself.

Cool, thanks for ignoring what I said to say the same thing. Just posting this as further proof to yourself of what you are doing

 

[Kisai]

9 hours ago, A51UK said:

Microsoft do not need to response as it is software company not hardware company. They got Windows 10 on ARM now, may make on sale for public if ARM CPU become more on desktop. I think Microsoft would or need to care very little about  Apple Desktop Silicon as it make basic no different to them. They can put Windows on basic on any CPU if need to or want to.  Apple Desktop Silicon is not a real bit deal. 

Microsoft can decide what CPU it wants Windows to work on, and thus far that has been "x86-64 compatible cpu's" plus "AARCH64" aka ARMv8-A.

 

However x86-64, most software is compiled to "essentially the minimum", which is essentially any system with at least DDR2, requires the CMPXCHG16B instruction for Windows 8.1 (aka 10) to install and operate.

 

There is no such animal as a "ARMv8-A" general purpose CPU that you can just pop into any motherboard with a CPU socket. Until that happens, there will be no widespread support for ARM. It's a chicken-and-egg problem. At this point it's probably not even desirable, as it would be leaving features in existing SoC ARM parts (eg nVidia) completely unusable.

 

There needs to be some kind of standard by which all the optional SoC parts can be initialized and utilized by any OS, and thus far, you'd think that would be the purpose of UEFI, but nope.

[LAwLz]

1 hour ago, leadeater said:

Quote

If you want to have a zen 3 core performs as well as a Firestorm core, then the core itself, not including IO or anything like that, will use about 15 watts of power.

Correct, literally nobody has said otherwise. However what is your point with this? This is were you seem to get lost in all of this because nobody has said otherwise. What actually lead up to the conversation about Zen 3 per core power efficiency were claims, not by you but you joined in, from people that Ryzen Mobile is or was going to be many hundreds of percent higher power usage which is simply an illogical and incorrect statement to say which is why I addressed those specific claims with the per core clock scaling power.

Here is a quote from you a couple of threads ago:

  

  

On 12/10/2020 at 11:05 PM, leadeater said:

On 12/10/2020 at 11:02 PM, LAwLz said:

On 12/10/2020 at 10:50 PM, leadeater said:

On 12/10/2020 at 10:31 PM, Lord Vile said:

POWER houses is the right word. The M1 runs 4 performance cores at less power than a Ryzen 3rd gen chip takes to run one.  

No it doesn't, that is factually not correct.

No, it is factually correct.

If you load the 5950X with a single thread, all other cores being shut off, the total package power of the 5950X will be 49 watts. The core power will be 22 watts.

If you load up the M1 with a multithreaded workload, all other cores being active, the total power of the entire Mac Mini computer, including fans, storage, memory, etc, is ~26 watts.

 

That statement is only untrue if you start doing math that includes a lot of asterixis where you start subtracting a lot of stuff from the Ryzen CPU and try to extrapolate what one core might use in an idea scenario, or if you downclock the Ryzen chip to be slower than the M1.

 

If we look at it as "how much power does my computer use if I load the M1 with four threads" vs "how much power does my computer use if I load the 5950X with a single thread" then that statement is true.

No it is not correct, been down this before.

 

 

 

In case the quoting is hard to understand, someone said a single Zen 3 core uses more power than the M1 running all 4 cores.

You said that was incorrect.

I said it was correct if you looked at the 5950X. The only time it wasn't correct was if you downclocked the cores to the point where it didn't perform as well as the Firestorm cores (like in the mobile SKUs), and you said I was wrong and @GDRRiley gave it a "funny" reaction as an insult.

The entire thread is full of you being rude, repeatedly saying "you're wrong" to people, and cherry picking numbers to get whichever results you want.

[leadeater]

2 hours ago, LAwLz said:

In case the quoting is hard to understand, someone said a single Zen 3 core uses more power than the M1 running all 4 cores.

You said that was incorrect.

I said it was correct if you looked at the 5950X. The only time it wasn't correct was if you downclocked the cores to the point where it didn't perform as well as the

Cool story, yes it is possible for a Zen 3 core to use more but that entirely depends on which CPU you look at. His implication and what he was trying to say, and you know it, is all Zen 3 CPUs would do that. I took issue with, and still do, with that statement and you picking a single CPU, a desktop one for that matter, that shows what you want to prove an inane rhetoric point which is the problem in and of itself.

 

I said it was false and continue to say it is false because I'm not stupid, I know exactly what his intent was and what he was implying, I bet you do too if you actually step back and have an objective look at his posts in that topic and other M1 topics. He and others are posting just the worst and most utterly silly trash talk possible with zero regard to truth and facts and only happen to chance on some semblance of truth occasionally more by chance than anything else, however like I said I'm not stupid I know what they are actually trying to say.

 

The rhetoric which in that very topic I was pointing to is THE problem. M1 is not tech gods gift to CPUs and simply isn't as groundbreaking as you want it to be, it does nothing special or different to anyone else, it just happens to be using later technology than anyone else is but that by no means makes it groundbreaking. It's no different than being the current fastest 100m runner, doesn't mean you actually have the current world record i.e, groundbreaking. However this running example uses speed as what is and is not groundbreaking where as with CPUs speed alone or power efficiency alone does not make something groundbreaking. Groundbreaking here would be the first multithread CPU, or the first multicore CPU, or the first multi chip CPU i.e. some actually new groundbreaking technology not more of the same.

 

If you want to continue to support "hurr durrr" go right ahead, leave me out.

 

Now that Zen 3 mobile parts actually exist we can stop doing this stupid dance and only using desktop parts because he and others refused to even look at Zen 2 mobile parts which of no shock to me, like I said would happen, use the same power. So now that it is possible how about we start comparing mobile parts to mobile parts? Seems reasonable does it not?

 

2 hours ago, LAwLz said:

and cherry picking numbers to get whichever results you want.

If you and others do the same I have zero issue also doing it, not that I'm actually saying I have been cherry picking. If you want to believe that's what I've been doing I'm actually totally fine with it. Don't complain to me about receiving in kind.

[Amias]

I gotta say. After spending about a 3 weeks now with my wifes M1 Macbook Air, I'm not too impressed to be honest, considering all the hype.

 

It's so far utterly crawled running OneDrive. Crashed and gotten lagging doing basic things, logging in, installing software. It's keyboard stopped working for half an hour, then suddenly started reworking. But more importantly I have a late 2013 13" macbook pro i5 ... and it just doesn't seem to blow it away. Some things crush my old laptop like youtube and playing videos and waking from sleep. But general internet browsing outside of high definition content feels no faster.

Even running photoshop, office suite and chrome feel pretty much on par with a my 7 year old macbook. Whether that's a testament to the old Macbook or not I'm not sure. It just feels like maybe a smidge faster than my 7 year old laptop ... Excel felt damn right laggy ...

 

I was expecting things like software to open instantly. an 80% hit to x86 hardware ... not slowing to a crawl. Hmmmmm.