Discussion Apple Silicon SoC thread

[Eug]

M1
5 nm
Unified memory architecture - LP-DDR4
16 billion transistors

8-core CPU

4 high-performance cores
192 KB instruction cache
128 KB data cache
Shared 12 MB L2 cache

4 high-efficiency cores
128 KB instruction cache
64 KB data cache
Shared 4 MB L2 cache
(Apple claims the 4 high-effiency cores alone perform like a dual-core Intel MacBook Air)

8-core iGPU (but there is a 7-core variant, likely with one inactive core)
128 execution units
Up to 24576 concurrent threads
2.6 Teraflops
82 Gigatexels/s
41 gigapixels/s

16-core neural engine
Secure Enclave
USB 4

Products:
$999 ($899 edu) 13" MacBook Air (fanless) - 18 hour video playback battery life
$699 Mac mini (with fan)
$1299 ($1199 edu) 13" MacBook Pro (with fan) - 20 hour video playback battery life

Memory options 8 GB and 16 GB. No 32 GB option (unless you go Intel).

It should be noted that the M1 chip in these three Macs is the same (aside from GPU core number). Basically, Apple is taking the same approach which these chips as they do the iPhones and iPads. Just one SKU (excluding the X variants), which is the same across all iDevices (aside from maybe slight clock speed differences occasionally).

EDIT:



M1 Pro 8-core CPU (6+2), 14-core GPU
M1 Pro 10-core CPU (8+2), 14-core GPU
M1 Pro 10-core CPU (8+2), 16-core GPU
M1 Max 10-core CPU (8+2), 24-core GPU
M1 Max 10-core CPU (8+2), 32-core GPU

M1 Pro and M1 Max discussion here:

Page 78 - Discussion - Apple Silicon SoC thread

Page 78 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M1 Ultra discussion here:

Page 109 - Discussion - Apple Silicon SoC thread

Page 109 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M2 discussion here:

Page 127 - Discussion - Apple Silicon SoC thread

Page 127 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Second Generation 5 nm
Unified memory architecture - LPDDR5, up to 24 GB and 100 GB/s
20 billion transistors

8-core CPU

4 high-performance cores
192 KB instruction cache
128 KB data cache
Shared 16 MB L2 cache

4 high-efficiency cores
128 KB instruction cache
64 KB data cache
Shared 4 MB L2 cache

10-core iGPU (but there is an 8-core variant)
3.6 Teraflops

16-core neural engine
Secure Enclave
USB 4

Hardware acceleration for 8K h.264, h.264, ProRes

M3 Family discussion here:

Page 215 - Discussion - Apple Silicon SoC thread

Page 215 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

 

[SpudLobby]

SMT is probably pointless in a world where you have big and little cores. How would it benefit Apple to add a second thread to its P cores when they have very capable E cores just sitting there, sipping power? That's where you want that thread to run. They aren't designing servers, they don't care as much about maximizing throughput for unlimited MT code as Intel/AMD do.

Isn't Intel dropping HT in its next gen CPUs? Right after they added their own capable E cores? I don't think that's a coincidence.

Exactly.

Yes the throughput gain is there on MT for very minimal (like ~5-10% die area and closer to lower end apparently) logic but validation is a PITA, security is an issue, and the gains vary by workload and are more meaningful for servers. Scheduling probably becomes lot more painful with P + E and SMT too. We’ll be better off without it in client. And that’s where things are headed.

Performant, efficient ST + area and extremely energy efficient little cores for low QoS tasks is the way to go for anything from a phone to a tablet, headset, laptop.

 

[eek2121]

I knew of nowhere else to post this, and I am behind on this thread, but I needed to share this somewhere with an okay amount of anonymity. I was approached by Apple recruiters 3X recently (last one was 2 hours ago) about building/implementing an LLM, so they are most certainly planning something. I know there have been rumors and such, but I haven't really been following along since I had hoped they'd focus on "on device" solutions, but with the details I was given, they are definitely planning something bigger, and I doubt, at least based on the details I was given, that they are partnering with anyone.

That being said, these details were not accompanying an NDA so take the info I provided for what it is worth. 2 of the jobs were pitched via direct Apple recruiters, today's offer was an indirect one. I turned all 3 of them down because I've exited the industry and I am becoming a full time consultant.

 

[poke01]

I knew of nowhere else to post this, and I am behind on this thread, but I needed to share this somewhere with an okay amount of anonymity. I was approached by Apple recruiters 3X recently (last one was 2 hours ago) about building/implementing an LLM, so they are most certainly planning something. I know there have been rumors and such, but I haven't really been following along since I had hoped they'd focus on "on device" solutions, but with the details I was given, they are definitely planning something bigger, and I doubt, at least based on the details I was given, that they are partnering with anyone.

That being said, these details were not accompanying an NDA so take the info I provided for what it is worth. 2 of the jobs were pitched via direct Apple recruiters, today's offer was an indirect one. I turned all 3 of them down because I've exited the industry and I am becoming a full time consultant.

So the rumours are that Apple will do the “on-device AI” themselves and cloud AI most likely for chatbots they will partner with google.

WWDC should give a more clear understanding of their moves

 

[eek2121]

Apple is doomed

Not really, You should maybe study economics before proclaiming this. Intel is still the #1 chip maker for a reason.

Please, dear god, stop making the chip in the phones faster for no reason. It's already so hot, the battery can only hold so much electro-juices. What if I didn't have to charge the damn thing every day?

Uh, just because you own an Android phone...

EDIT: even with my 14th gen iPhone, I do not have to charge every day, and my phone is used 8+ hours a day. With that usage is lasts nearly 2 days without needing a recharge.

Zen5 is only just widening their core from the 4 wide design of Zen1 though.

They still have a lot of open road to explore before they are tapped out.

It is a lot greater than 40%. Look at a certain previous release, which may or may not be relevant to current ones....

Let us discuss about Apple's OLED Macbooks here👇

Question - 👑Apple: The OLED Roadmap👑

Apple will launch the 11- and 12.9-inch iPad Pro in 2024. Apple will launch an 8.3-inch OLED iPad mini tablet PC in 2026 and a 10.8-inch OLED iPad Air tablet PC with an LTPS backplane and single-structure RGB OLED in 2027. Omdia estimates that BOE plans to produce these panels in its newly...

forums.anandtech.com

I am so happy the current OLED rumors are confirming this. Apple has no idea how much money I'm about to throw at them as a result...unless it is false (not likely).

The next gen chips won't be as amazing as most people are expecting, but they will be great.

 

[igor_kavinski]

Apple has no idea how much money I'm about to throw at them as a result

So you are fully prepared for the early adopter issues, cursing at them and feeling like you wasted your money but still convincing yourself that it was the best decision you ever made because Apple?

 

[poke01]

So you are fully prepared for the early adopter issues, cursing at them and feeling like you wasted your money but still convincing yourself that it was the best decision you ever made because Apple?

For their new new products like Vision Pro I would wait for Gen 3 or Gen 4.

For their new generation of OLED MacBooks, should be fine as they are trialing the OLED this year in the iPads before they put them into MacBooks.

That kinda should tell how new the oxide OLED tech is. It’s bound to have some issues on first gen.

 

[Doug S]

So the rumours are that Apple will do the “on-device AI” themselves and cloud AI most likely for chatbots they will partner with google.

WWDC should give a more clear understanding of their moves

I would read that Google talk as 100% about AI search. They want to keep everything on device to the extent possible, but obviously you can't do internet search on device so they want to see if Google will write them even more ridiculously large checks to be the default for AI enhanced internet search.

 

[SteinFG]

they want to see if Google will write them even more ridiculously large checks to be the default for AI enhanced internet search.

Why would google spend more money to spend more money?

 

[igor_kavinski]

Why would google spend more money to spend more money?

Because they don't want their users feeling all left out while Windows users' and their kids are gloating about Co-pilot.

 

[FlameTail]

Because they don't want their users feeling all left out while Windows users' and their kids are gloating about Co-pilot.

Gemini vs ChatGPT

Google vs Microsoft

Godzilla vs Kong

 

[Glo.]

Gemini vs ChatGPT

Google vs Microsoft

Godzilla vs Kong

Aliens vs Goblins.

 

[FlameTail]

Apple M4's CPU will be the biggest jump in M series history, since the M1. It will feature the new Quasar P-core, which brings a colossal 25% performance improvement (+30% IPC and clock regression). The first clean sheet core design since Firestorm. The Pulsar E-core, is a more modest iteration of Sawtooth, with it's focus mainly on efficiency. These improvements, combined with the new 4P+6E CPU core configuration, will give the M4 a whopping 40% improvement in multi-core performance.

Source: My fantasy

 

[okoroezenwa]

Quite the wish lol. I would, though, like to see a double digit IPC increase and I do hope the great improvements they’ve been making with their E-cores continues.

GPU is where I’d love to see greater gains though.

 

[Doug S]

Why would google spend more money to spend more money?

Google profits off the search deal with Apple via their search ad business. Do you think that AI search won't have the same business model? That's the only business Google knows. Every business they run that's not advertising collectively loses money, and that's always been the case for them.

 

[igor_kavinski]

oops wrong thread

 

[poke01]

Apple Updates App Store Guidelines to Permit Game Emulators, Website Links in EU Music Apps

Apple today updated its App Store guidelines to comply with an anti-steering mandate levied by the European Commission. Music streaming apps like Spotify are now permitted to include a link or buy button that leads to a website with information about alternative music purchasing options, though...

forums.macrumors.com

"Game emulators have managed to sneak onto the App Store several times over the years by using hidden functionality, but Apple has not explicitly permitted them until now. The rule change that allows for game emulators is worldwide, as is support for apps that offer mini apps and mini games."

game Emulators are allowed now on App Store

 

[igor_kavinski]

game Emulators are allowed now on App Store

Nintendo lawyers will surely rejoice at the new opportunity to sue someone

 

[Apokalupt0]

Apple M4's CPU will be the biggest jump in M series history, since the M1. It will feature the new Quasar P-core, which brings a colossal 25% performance improvement (+30% IPC and clock regression). The first clean sheet core design since Firestorm. The Pulsar E-core, is a more modest iteration of Sawtooth, with it's focus mainly on efficiency. These improvements, combined with the new 4P+6E CPU core configuration, will give the M4 a whopping 40% improvement in multi-core performance.

Source: My fantasy

Do we have any solid source on what the expected perf increase is with M4 and A18 Pro?
Other than random Geekbench 6 leaks which are hard to believe?

 

[okoroezenwa]

Nope.

 

[GC2:CS]

Do we have any solid source on what the expected perf increase is with M4 and A18 Pro?
Other than random Geekbench 6 leaks which are hard to believe?

Twitter acount @tech_reve posted some scores. +25% single and multi for the A18 Pro vs A17 Pro
+35% GPU as well.
They supposedly predicted A17 correctly but this time got taken down immediately.

Also A18 2P+5E+5G A18 Pro 2P+6E+6G.
I stikl think they should sacrifice a P core for the cheaper version.

 

[FlameTail]

Twitter acount @tech_reve posted some scores. +25% single and multi for the A18 Pro vs A17 Pro
+35% GPU as well.
They supposedly predicted A17 correctly but this time got taken down immediately.

I don't recall him leaking this.

Also A18 2P+5E+5G A18 Pro 2P+6E+6G.

6G???

I stikl think they should sacrifice a P core for the cheaper version.

Nope. It better be;
A18 Pro : 2P+6E
A18 : 2P+4E

 

[ashFTW]

I don't recall him leaking this.

6G???

Nope. It better be;
A18 Pro : 2P+6E
A18 : 2P+4E

6 graphics cores, not 6G wireless.

 

[poke01]

Macs lead global PC growth as shipments return to pre-pandemic levels - 9to5Mac

IDC is out with a fresh report today on global PC shipments for Q1 2024. While Apple still doesn’t lead...

9to5mac.com

Looks like those M3 macbooks helped with sales for Apple.

 

[igor_kavinski]

Macs lead global PC growth as shipments return to pre-pandemic levels - 9to5Mac

IDC is out with a fresh report today on global PC shipments for Q1 2024. While Apple still doesn’t lead...

9to5mac.com

Looks like those M3 macbooks helped with sales for Apple.

Without a breakdown by chip type, we can't be sure. What does M3 bring to the table that would cause a sudden uptick in uptake by consumers? Hardware raytracing in the GPU? Are Mac users now running the 3DMark Solar Bay raytracing benchmark on their Macs? Are Apple stores wooing customers by showing the Macs running this benchmark? Are Apple stores/geniuses actively demoing games on the new Macs? Anyone have concrete answers?

 

[poke01]

Without a breakdown by chip type, we can't be sure. What does M3 bring to the table that would cause a sudden uptick in uptake by consumers? Hardware raytracing in the GPU? Are Mac users now running the 3DMark Solar Bay raytracing benchmark on their Macs? Are Apple stores wooing customers by showing the Macs running this benchmark? Are Apple stores/geniuses actively demoing games on the new Macs? Anyone have concrete answers?

It could be users from M1 or older Intel Macs and some might be new customers.

The M3 series GPU is big step up from M1. Blender 4 does support M3 hardware acceleration of Ray tracing which brings huge speed increases compared to M1/2.

The M3 Max CPU is the best mobile CPU out there. Maybe software companies for code compiling? They like to hand out laptops.
Reviewers also highlighted the M3 Max too that may caused an uptick of users buying M3 Max instead of M3 Pro.

Other than that no major improvements. This is just my speculation.