Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

bradman83

macrumors 65816
Oct 29, 2020
1,277
3,241
Buffalo, NY
Wait do we know M1 pro and above is using A15 arch?
All M1-series chips use the same Firestorm and Icestorm core microarchitectures as the A14. The A15 uses newer Blizzard and Avalanche core designs. These might be the basis for the M2, or the M2 could be based on whatever weather phenomenon core codename they give the A16 depending on timing, but the M1 Pro/Max/Ultra still use the A14 cores.

Proof: Their single core benchmarks are all almost identical.
 

jeffpeng

macrumors regular
Aug 9, 2021
227
359
Both the A14 and the M1 have the same encoder hardware for ProRes. The A15 has newer, better and bigger encoder hardware. A lot of the additional silicon budget of the A15 went into the encoder and decoder blocks. Hence the A15 is faster at encoding and decoding ProRes.

This has (almost) nothing to do with how these chips performs otherwise. ProRes is rendered exclusively on those blocks, and those block do ProRes and only ProRes. This is what you'd call fixed function hardware, as it can't be utilized to do literally anything else. Where other general purpose computing blocks, such as GPU and CPU can run arbitrary code, these parts have the software more or less baked in. The advantage is that they are much faster and way more efficient at what they do. The disadvantage is that they, as I said, can do only that, and that there is no updating the algorithms with later versions as it is what it is and cannot be changed.

This is also the reason why Apple Silicon flies rendering these codecs, but falls relatively flat rendering other codecs it has no specific hardware for. This is why the M1 chips are only really good for video editing if you commit to using Apples proprietary codecs, but if you do they are very, very fast at doing that.
 

macduke

macrumors G5
Jun 27, 2007
13,468
20,530
Make the video. It’s even BETTER that it’s inconclusive because you can ask folks to “drop why they think this is in the comments” and you’ll get good engagement numbers right off the bat. Then, you have a reason to ask people to subscribe for part 2. :)
This mans YouTubes.

Though I will offer just a teeny-tiny small correction: Don’t just ask them to subscribe, ask them to smash that like button too. And don’t forget to ring the notification bell. And if they want even more content, they can check out your Patreon, or they can buy your merch. Maybe sign up for your crypto scam or simp for you on OnlyFans. Sell tickets to your live show or boxing event. Sponsored by Blue Apron/Master Class/Me Undies/SquareSpace/Raid Shadow Legends. [Insert endearing “He said the thing!” catch-phrase sign-off here.]

Also make sure to say something controversial so all the tea/drama channels pick up on it and you go viral. Even better if it’s something hurtful or hypocritical about a fellow YouTuber. Make sure to be simultaneously too woke about some things and not woke enough about others. It’s important to irritate both extremes to drive engagement.
 

jeffpeng

macrumors regular
Aug 9, 2021
227
359
This mans YouTubes.

Though I will offer just a teeny-tiny small correction: Don’t just ask them to subscribe, ask them to smash that like button too. And don’t forget to ring the notification bell. And if they want even more content, they can check out your Patreon, or they can buy your merch. Maybe sign up for your crypto scam or simp for you on OnlyFans. Sell tickets to your live show or boxing event. Sponsored by Blue Apron/Master Class/Me Undies/SquareSpace/Raid Shadow Legends. [Insert endearing “He said the thing!” catch-phrase sign-off here.]

Also make sure to say something controversial so all the tea/drama channels pick up on it and you go viral. Even better if it’s something hurtful or hypocritical about a fellow YouTuber. Make sure to be simultaneously too woke about some things and not woke enough about others. It’s important to irritate both extremes to drive engagement.
Alright. Now I know everything to become a tuber. Just quit my day job. Thank you so much ?
 
  • Like
Reactions: macduke

ian87w

macrumors G3
Feb 22, 2020
8,704
12,638
Indonesia
Yes, you heard right.

I just bought the iPad Pro M1 8 GB RAM and 11 inch. I decided to test how faster the M1-chip was, compared to my iPhone 13 Pro Max, and therefore downloaded LumaFusion. And now it's become a little weird. Because the A15 chip is actually faster than the M1 chip.



The original file was an: 8K IPB 29,97 fps MPEG-4 film with a duration of 1 minute and 54 seconds.

The output file was 4K 29,97 fps Apple ProRess 422 with the format Quicktime (.mov)

A15 chip export time: 2 minutes and 20 seconds
M1 chip export time: 3 minutes and 32 seconds

How is that possible. I mean, the M1 chip has a much higher Multi-core score, so why is the iPhone still faster in rendering/exporting the file?

I ran the test three times, and all with the same result.

Actually, I am beginning a new YouTube-channel, and wanted to make a video comparison with the iPad Pro M1, Mac Mini M1 (16 GB RAM), MacBook Air M1 (8 GB RAM), and a monster of a PC - an Nvidia Geforce RTX 3090 with Intel Core i9 12. gen - all that together and comparing multiple rendering processes and see how much different they all are

But I am not sure about making the video, because I cant figure out, why the M1-chip is slower than a A15 chip

Is there anyone who can explain that?
A15 has newer cores than M1 as M1 is based on A14 cores.
The advantage of your M1 iPad Pro will be in sustained performance as the iPad has better thermal headroom than the iPhone. Of course, encoding a short video won't show that advantage. But I believe once you are encoding longer videos, the iPhone will probably throttle sooner than the iPad.

Short burst of tests will not fully stress the systems, thus the newer cores of the A15 wins.
 

ikir

macrumors 68020
Sep 26, 2007
2,175
2,366
Yes, you heard right.

I just bought the iPad Pro M1 8 GB RAM and 11 inch. I decided to test how faster the M1-chip was, compared to my iPhone 13 Pro Max, and therefore downloaded LumaFusion. And now it's become a little weird. Because the A15 chip is actually faster than the M1 chip.



The original file was an: 8K IPB 29,97 fps MPEG-4 film with a duration of 1 minute and 54 seconds.

The output file was 4K 29,97 fps Apple ProRess 422 with the format Quicktime (.mov)

A15 chip export time: 2 minutes and 20 seconds
M1 chip export time: 3 minutes and 32 seconds

How is that possible. I mean, the M1 chip has a much higher Multi-core score, so why is the iPhone still faster in rendering/exporting the file?

I ran the test three times, and all with the same result.

Actually, I am beginning a new YouTube-channel, and wanted to make a video comparison with the iPad Pro M1, Mac Mini M1 (16 GB RAM), MacBook Air M1 (8 GB RAM), and a monster of a PC - an Nvidia Geforce RTX 3090 with Intel Core i9 12. gen - all that together and comparing multiple rendering processes and see how much different they all are

But I am not sure about making the video, because I cant figure out, why the M1-chip is slower than a A15 chip

Is there anyone who can explain that?
So you are cherry picking only the case? it could be due to better new generation encoder. A15 is not faster than M1 speaking in general.
 

jeffpeng

macrumors regular
Aug 9, 2021
227
359
So you are cherry picking only the case? it could be due to better new generation encoder. A15 is not faster than M1 speaking in general.
It's a very valid point he's making.

Unless you can actually utilize the additional CPU- and GPU-Cores of the M1 the A15 is the faster chip, and this is an excellent example of it. That's also why the iPad Air would have been better off with a A15 instead of the M1 and that upgrade is such a nothingburger. There hardly is any application that takes actual advantage of the aforementioned features of the M1 on the iPad. Technically the better GPU would offer better graphics, but I doubt any game will actually be sensitive to that.

Since it appears as if Apple has stopped producing the A14 I fully expect the next generation iPad Nothing to also feature the M1 simply for economy of scale reasons. Then maybe we'll see third party applications take advantage of the M1 on the iPad eventually.
 
  • Like
Reactions: G5isAlive

Sankew

Suspended
Dec 18, 2020
38
57
Yes, you heard right.

I just bought the iPad Pro M1 8 GB RAM and 11 inch. I decided to test how faster the M1-chip was, compared to my iPhone 13 Pro Max, and therefore downloaded LumaFusion. And now it's become a little weird. Because the A15 chip is actually faster than the M1 chip.



The original file was an: 8K IPB 29,97 fps MPEG-4 film with a duration of 1 minute and 54 seconds.

The output file was 4K 29,97 fps Apple ProRess 422 with the format Quicktime (.mov)

A15 chip export time: 2 minutes and 20 seconds
M1 chip export time: 3 minutes and 32 seconds

How is that possible. I mean, the M1 chip has a much higher Multi-core score, so why is the iPhone still faster in rendering/exporting the file?

I ran the test three times, and all with the same result.

Actually, I am beginning a new YouTube-channel, and wanted to make a video comparison with the iPad Pro M1, Mac Mini M1 (16 GB RAM), MacBook Air M1 (8 GB RAM), and a monster of a PC - an Nvidia Geforce RTX 3090 with Intel Core i9 12. gen - all that together and comparing multiple rendering processes and see how much different they all are

But I am not sure about making the video, because I cant figure out, why the M1-chip is slower than a A15 chip

Is there anyone who can explain that?
Okay so the M1 has no ProRes Encoders in it, so in a ProRes Export it will be slower compared to the iPhone 13
 
  • Like
Reactions: ikir

Sankew

Suspended
Dec 18, 2020
38
57
I rarely say this:

Architecture:

A14 -> M1
A14 -> A15
A15 -> M1 Pro/Max
M1 Max -> M1 Ultra

Encoders/Decoders/ISP/Audio Engine:

A15 (iPhone 13) -> M1 Pro
A15 (iPhone Pro) -> M1 Max
the M1 is basically a higher clocked version of A14 along with the extra CPU and GPU cores
A14 was clocked at 3.0Ghz whereas the M1 had it at 3.2Ghz
the A15 is just the A14 with the cores clocked to 3.2Ghz and the ProRes Encoders added to it for ProRes Recording
GeekBench 5 Single Core Results between the A14,A15 and the M1 make some sense in this regard
1648550767244.png
 

MauiPa

macrumors 68040
Apr 18, 2018
3,438
5,084
I got this from the lumafusion website:

What devices will LumaFusion work with?​

LumaFusion can be installed on any iOS device that is using iOS 15.0 or later. This means that LumaFusion can be installed on most iPads, iPadPros, iPhones and even the latest iPod.

Additionally, LumaFusion can also be installed on a Mac that is using the M1 processor. This type of Mac allows you to install iOS apps and use them

So basically lumafusion is an iOS app. Still looking
 

Macative

Suspended
Mar 7, 2022
834
1,319
"I'm starting a YouTube channel to talk about these things, but I don't understand any of them, so can someone please tell me so I can start my YouTube channel, thanks."

LOL
 
  • Haha
Reactions: G5isAlive

headlessmike

macrumors 65816
May 16, 2017
1,432
2,831
the A15 is just the A14 with the cores clocked to 3.2Ghz and the ProRes Encoders added to it for ProRes Recording
According to Anandtech's review of the A15 there's a lot more to it than just a higher clocked A14 with extra encoders. The performance cores received a noticeable increase in performance, the efficiency cores saw a major overhaul, and on top of that the cache was doubled. This reduced the energy consumption of the A15 to below that of the A14 despite it running at a higher clock speed.
 

Unregistered 4U

macrumors G4
Jul 22, 2002
10,597
8,591
A15 chip export time: 2 minutes and 20 seconds
M1 chip export time: 3 minutes and 32 seconds
I’ve been thinking about this a bit this morning and I’m curious as to what an Intel MacBook Air would have as an export duration. Because, while the M1 is not as fast as the A15, the stuff the M1 can do related to ProRes is still impressive. One wouldn’t expect the same performance from a MacBook Air prior to Apple Silicon.

For all the limitations of Apple devices, there’s a wealth of base performance… the phone is more performant than the MacBook, sure, but both outperform the vast majority of PC’s. That’s just crazy :)
 

Unregistered 4U

macrumors G4
Jul 22, 2002
10,597
8,591
Both the A14 and the M1 have the same encoder hardware for ProRes. The A15 has newer, better and bigger encoder hardware. A lot of the additional silicon budget of the A15 went into the encoder and decoder blocks. Hence the A15 is faster at encoding and decoding ProRes.
This is what I thought.
 
  • Like
Reactions: AmazingTechGeek

JPack

macrumors G5
Mar 27, 2017
13,507
26,138
It's a very valid point he's making.

Unless you can actually utilize the additional CPU- and GPU-Cores of the M1 the A15 is the faster chip, and this is an excellent example of it. That's also why the iPad Air would have been better off with a A15 instead of the M1 and that upgrade is such a nothingburger. There hardly is any application that takes actual advantage of the aforementioned features of the M1 on the iPad. Technically the better GPU would offer better graphics, but I doubt any game will actually be sensitive to that.

Since it appears as if Apple has stopped producing the A14 I fully expect the next generation iPad Nothing to also feature the M1 simply for economy of scale reasons. Then maybe we'll see third party applications take advantage of the M1 on the iPad eventually.

The A-series processors were I/O limited and didn't have full external display support or the bandwidth for moving large files. It helps round out the weaknesses of iPad, of course the OS now being the weak point.
 

AmazingTechGeek

macrumors 6502a
Mar 6, 2015
685
304
Los Angeles
the M1 is basically a higher clocked version of A14 along with the extra CPU and GPU cores
A14 was clocked at 3.0Ghz whereas the M1 had it at 3.2Ghz
the A15 is just the A14 with the cores clocked to 3.2Ghz and the ProRes Encoders added to it for ProRes Recording
GeekBench 5 Single Core Results between the A14,A15 and the M1 make some sense in this regard
View attachment 1982965
Yup. Just small iterative changes for now. I expect M2 and A16 to be architecturally different from M1 series.
 

jeffpeng

macrumors regular
Aug 9, 2021
227
359
The A-series processors were I/O limited and didn't have full external display support or the bandwidth for moving large files. It helps round out the weaknesses of iPad, of course the OS now being the weak point.
That's a valid point. Although, frankly, I feel this is also an oversight on Apples part. The A15 should already have included at least USB4 capabilities. But point taken. For this specific use case the change makes sense.
 

jeffpeng

macrumors regular
Aug 9, 2021
227
359
Yup. Just small iterative changes for now. I expect M2 and A16 to be architecturally different from M1 series.
The biggest jump chips get is from processes, and as it appears M2/A16 will be on TSMC N4, which is technically N5++, not the new N3 node, which seems to be delayed. So while I guess improvements on the architecture side can matter (and they do with the A15, just not sooo much with the performance CPU cores, but everything else is notably better), don't expect a revolutionary improvement, but more iterative gains.

Personally I don't think the M2 will not outright obsolete the M1 chips so much you will "need" to upgrade. Some will want to, but these some always get the latest stuff. And honestly I think that's okay. The M1 chips are crazy fast and efficient as they are, and you will be fine daily driving those past 2025.
 
  • Like
Reactions: sorgo †

iKnowMyABGz

macrumors newbie
Apr 1, 2022
9
5
Honolulu, HI
Yes, you heard right.

I just bought the iPad Pro M1 8 GB RAM and 11 inch. I decided to test how faster the M1-chip was, compared to my iPhone 13 Pro Max, and therefore downloaded LumaFusion. And now it's become a little weird. Because the A15 chip is actually faster than the M1 chip.



The original file was an: 8K IPB 29,97 fps MPEG-4 film with a duration of 1 minute and 54 seconds.

The output file was 4K 29,97 fps Apple ProRess 422 with the format Quicktime (.mov)

A15 chip export time: 2 minutes and 20 seconds
M1 chip export time: 3 minutes and 32 seconds

How is that possible. I mean, the M1 chip has a much higher Multi-core score, so why is the iPhone still faster in rendering/exporting the file?

I ran the test three times, and all with the same result.

Actually, I am beginning a new YouTube-channel, and wanted to make a video comparison with the iPad Pro M1, Mac Mini M1 (16 GB RAM), MacBook Air M1 (8 GB RAM), and a monster of a PC - an Nvidia Geforce RTX 3090 with Intel Core i9 12. gen - all that together and comparing multiple rendering processes and see how much different they all are

But I am not sure about making the video, because I cant figure out, why the M1-chip is slower than a A15 chip

Is there anyone who can explain that?
I wonder if they will ever adopt cool OS naming like android and maybe name its new OS something like AppleCider? ¯\_(ツ)_/¯ and for the Pro Models AppleCider-V. Get it, V bc people use apple cider vinegar to do gods work!
 
  • Haha
Reactions: sorgo †

sorgo †

Cancelled
Feb 16, 2016
2,870
7,046
I wonder if they will ever adopt cool OS naming like android and maybe name its new OS something like AppleCider? ¯\_(ツ)_/¯ and for the Pro Models AppleCider-V. Get it, V bc people use apple cider vinegar to do gods work!
And rename HomePod software to “AppleFritter,” because they’re always getting fried
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.