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

M4 Family discussion here:

Page 263 - Discussion - Apple Silicon SoC thread

Page 263 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

 

[Eug]

I cancelled my order it was an impulse purchase. I already have a great desktop which I upgraded this year. Don’t need another.

I rather save up

I don't need one either, but I'm buying the Mac mini anyway. I think/hope I can get a reasonable amount selling the existing M1 / 16 GB / 1 TB Mac mini.

BTW, I just got a new internet modem with 2.5 Gbps Ethernet port, and rejigged the cabling so I now have a direct run from the modem in the basement to the home office two floors up, bypassing all the existing Gigabit switches. Hopefully my existing CAT5e will be able to handle a 2.5 Gbps link for the new Mac mini.

 

[The Hardcard]

I just came across these 13" iPad Pro M4 Geekbench 6.3 scores:

4005 / 15006
4001 / 15007

Both on 18.2. Another performance enhancing update? It would be nice to see update comparisons on non benchmark code.

 

[poke01]

I don't need one either, but I'm buying the Mac mini anyway. I think/hope I can get a reasonable amount selling the existing M1 / 16 GB / 1 TB Mac mini.

BTW, I just got a new internet modem with 2.5 Gbps Ethernet port, and rejigged the cabling so I now have a direct run from the modem in the basement to the home office two floors up, bypassing all the existing Gigabit switches. Hopefully my existing CAT5e will be able to handle a 2.5 Gbps link for the new Mac mini.

Oh I'll still get it at some point next year but I need to save up for the Japan trip. Those $740AUD are better spent there.

At least you can bench your Mac mini

 

[jdubs03]

Oh I'll still get it at some point next year but I need to save up for the Japan trip. Those $740AUD are better spent there.

At least you can bench your Mac mini

I’m hoping Black Friday will present an opportunity for a few hundred USD off.

 

[poke01]

@FlameTail This is good article on Apple's NPU.

More Core, More Power: The Apple M4 Pro Advantage

Discover Apple's M4 Pro and M4 Max chips: improved core design, AI processing, and enhanced battery life delivering up to 24 hours in the new MacBook Pro.

creativestrategies.com

 

[mvprod123]

Here are the results of some tests from the full russian review of the MacBook Pro with a regular M4. Some comparisons are incorrect because the fanless MacBook Air is used for comparison. Interestingly, the reviewer said that M4 in After Effects is as fast as M1 Max.

 

[Mopetar]

My M1 MBP is still more than good enough so I see little reason to upgrade. The new Mac Mini is rather tempting at $500 just to hook up to the television.

I'm giving up on Apple making an iMac worth buying anytime soon though. I'll wait around to see what the M4 Studio upgrade looks like, but I don't get why they're allergic to releasing a 27" iMac with a Pro/Max M-series chip.

M4 Ultra should be sexy though.

 

[Glo.]

I have to tell you guys.

IMO, M4 Mac Mini and M4 MacBook Pro with 24 GB of RAM, and support for 3 displays are the best trading/financial, technical analysis machines for very little money. Especially the Mini.

 

[poke01]

Here are the results of some tests from the full russian review of the MacBook Pro with a regular M4. Some comparisons are incorrect because the fanless MacBook Air is used for comparison. Interestingly, the reviewer said that M4 in After Effects is as fast as M1 Max.
View attachment 110863View attachment 110864View attachment 110865

Speedometer 3.0 finally gets an increase with M4. Theres was barely any increase from M1 to M3.

 

[poke01]

Geekbench 5
M4 Pro

Mac16,11 - Geekbench

Benchmark results for a Mac16,11 with an Apple M4 Pro processor.

browser.geekbench.com

9950X

Gigabyte Technology Co., Ltd. X670E AORUS MASTER - Geekbench

Benchmark results for a Gigabyte Technology Co., Ltd. X670E AORUS MASTER with an AMD Ryzen 9 9950X processor.

browser.geekbench.com

 

[Eug]

Geekbench 5
M4 Pro

Mac16,11 - Geekbench

Benchmark results for a Mac16,11 with an Apple M4 Pro processor.

browser.geekbench.com

9950X

Gigabyte Technology Co., Ltd. X670E AORUS MASTER - Geekbench

Benchmark results for a Gigabyte Technology Co., Ltd. X670E AORUS MASTER with an AMD Ryzen 9 9950X processor.

browser.geekbench.com

Ah Geekbench 5. Very similar scores to each other.

M4 Pro - 2737 / 22872
9950X - 2702 / 26974

 

[igor_kavinski]

Ah Geekbench 5. Very similar scores to each other.

M4 Pro - 2737 / 26872
9950X - 2702 / 26974

Comparison shows that in ST, Zen 5 is mainly able to hold its own coz of two tests, namely AES and Machine Learning, probably thanks to AVX-512: https://browser.geekbench.com/v5/cpu/compare/22775563?baseline=23018554

Weirdly, in AES MT, Zen 5 just plain crashes HARD and can't perform as well in Machine learning so the situation is reversed. The other tests are more in Zen 5's favor thanks to better MT scaling. I would love to compare some SMT-less 9950X GB5 score with M4.

 

[FlameTail]

Ah Geekbench 5. Very similar scores to each other.

M4 Pro - 2737 / 26872
9950X - 2702 / 26974

M4 Pro is 22872

 

[igor_kavinski]

M4 Pro 14-cores giving the M1 Ultra 20-cores some tough time: https://browser.geekbench.com/v5/cpu/compare/23024218?baseline=23018554

 

[igor_kavinski]

Very curious result in MT, 10 cores M4 vs. 10 cores M1 Max: https://browser.geekbench.com/v5/cpu/compare/23024184?baseline=23023916

Not sure why MT performance would regress in some of the subtests at the same core count...

 

[The Hardcard]

Very curious result in MT, 10 cores M4 vs. 10 cores M1 Max: https://browser.geekbench.com/v5/cpu/compare/23024184?baseline=23023916

Not sure why MT performance would regress in some of the subtests at the same core count...

4P + 6E vs 8P + 2E. I’m impressed that MT is that close.

 

[igor_kavinski]

14900K hackintosh vs. M4 Pro: https://browser.geekbench.com/v5/cpu/compare/23023862?baseline=23018554

Wonder how long Apple will allow some users to "enjoy" the MacOS experience on the cheap.

 

[okoroezenwa]

14900K hackintosh vs. M4 Pro: https://browser.geekbench.com/v5/cpu/compare/23023862?baseline=23018554

Wonder how long Apple will allow some users to "enjoy" the MacOS experience on the cheap.

What do you mean? An M4 Pro machine is not cheap.

 

[Nothingness]

Very curious result in MT, 10 cores M4 vs. 10 cores M1 Max: https://browser.geekbench.com/v5/cpu/compare/23024184?baseline=23023916

Not sure why MT performance would regress in some of the subtests at the same core count...

Isn't the base you chose for M4 way below other M4?

Edit: my bad that is GB5. Doesn't show on mobile.

 

[igor_kavinski]

What do you mean? An M4 Pro machine is not cheap.

14900K, especially used, is comparatively cheap. Check the OS running on the 14900K

 

[Eug]

M4 Pro is 22872

Yes thx, fixed. I got confused with M4 Max which would likely be closer to 9950X for multicore.

 

[MS_AT]

Comparison shows that in ST, Zen 5 is mainly able to hold its own coz of two tests, namely AES and Machine Learning, probably thanks to AVX-512: https://browser.geekbench.com/v5/cpu/compare/22775563?baseline=23018554

Weirdly, in AES MT, Zen 5 just plain crashes HARD and can't perform as well in Machine learning so the situation is reversed. The other tests are more in Zen 5's favor thanks to better MT scaling. I would love to compare some SMT-less 9950X GB5 score with M4.

Didnt someone on this forum prove that AVX512 gives Zen5 no benefits in Geekbench?

 

[The Hardcard]

14900K hackintosh vs. M4 Pro: https://browser.geekbench.com/v5/cpu/compare/23023862?baseline=23018554

Wonder how long Apple will allow some users to "enjoy" the MacOS experience on the cheap.

I think this is likely as good as it’s going to get pricewise. I was surprised at how few price increases there were with the move to 16 GB base. I had been arguing against raising the base memory because I figured Apple would sell them for the same 16 GB prices.

Plus, their current path is working for them. Unless there is a world economic collapse, they are going to sell high volumes of these Macs. Not only do they take the performance crown in nearly every catagory, but several Mac SKUs offer features not possible elsewhere.

 

[Eug]

I think this is likely as good as it’s going to get pricewise. I was surprised at how few price increases there were with the move to 16 GB base. I had been arguing against raising the base memory because I figured Apple would sell them for the same 16 GB prices.

Plus, their current path is working for them. Unless there is a world economic collapse, they are going to sell high volumes of these Macs. Not only do they take the performance crown in nearly every catagory, but several Mac SKUs offer features not possible elsewhere.

I was guessing a $100 price increase for M4 16 GB vs M2 8 GB, with the old model still continuing to be sold for less, and I was thinking 3 USB ports for the M4. I’m very happy to be wrong on both counts.

I’ve been pairing my M1 1 TB Mac mini with a USB-C 10 Mbps 2 TB DRAM-less SSD, but for the new M4 512 GB Mac mini, I may get a new USB 4 40 GBPs 4 TB SSD with DRAM.

 

[Nothingness]

Didnt someone on this forum prove that AVX512 gives Zen5 no benefits in Geekbench?

It was on Zen4 IIRC.