WWDC 2023 Thread

[Jimmyjames]

I listened to the latest episode of ATP, so that you don't have to. The main reason to listen to this show is John Siracusa's epic rants, with Marco Arment as the "straight man", and Casey Liss as topic moderator. Siracusa's toaster reviews are legendary. That being said, he's hitting "old man yells at cloud" territory.

Regarding the Mac, they spent about five minutes on the new Mac Studio, and are happy that the sound issue appears to be fixed. Siracusa then spent a half hour ranting about how much he hates the new Mac Pro and it's not what he wants. Marco Arment points out that Siracusa purchased a 2019 Mac Pro with the sole intention of using it as a gaming PC. (Siracusa has stated that he only plays Destiny in Boot Camp.) Speaking of games, Siracusa rants for another half hour about how much he hates Apple's new gaming efforts and that they're going to fail. Then, they spend two minutes on Sonoma, and Siracusa is happy that they fixed his two pet bugs. Finally, Apple execs say that they saw the "Believe" Mac Pro shirts from ATP.

Aside from the bellyaching, they do point out something that I forgot to mention. When comparing the sound levels between the M1 Mac Studio and M2 Mac Studio, Apple's stats claim that, while idle or web browsing, the M1 Mac Studio is 15dB, compared to the M2 Mac Studio at 6dB. Decibels are logarithmic, as seen on this chart, which means that the M2 Mac Studio is approximately ~8 times quieter than the previous generation. That's a notable decrease in sound output at idle, therefore Apple definitely tinkered with the cooling solution.

Yeah, I knew it was gonna be bad and I therefore didn’t listen to it. I’m not sure what happened to Siracusa but I don’t like it! He has gone from outright curmudgeon to misrepresentation. If you look at the show notes for this episode, the benchmarks he provides are cherry-picked to paint Apple Silicon in the worst possible light. He links to gfxbench aztec normal tier to show the 4080 beating the Ultra. Normal tier is seldom used because given how easy most gpus deal with it, very tiny differences in drivers can have a significant impact. And really what does it matter if you get 800fps vs 1000fps? They’re both super high. “Mysteriously” he avoided the 4K Aztec off screen which shows the Ultra matching the 4080. I know he knows this information, because it was pointed out to him by me the other day. He then compares the Ultra to older AMD GPUs in Geekbench to paint the picture that Apple Silicon can’t match old gpus, carefully side stepping that once again, the Ultra is close to the 4080 and beats the 3090. I just think he’s got such an obsession with discreet gpus, he’s being dishonest now.

As for the gaming take, I am baffled how anyone can come to the conclusions he does. He is not alone. As we discussed previously, a certain group of old time Mac users are obsessed with proving that the Mac can’t game. On the other hand, there are some more sensible people out there. Whether it’s Gruber (I know), Guy English (a great and knowledgeable person in the community), Christina Warren etc.

I will just avoid the negative ones from now. Especially when I feel they are not debating in good faith.

 

[B01L]

oh no. Apple's first gen products always suck in longivity. The first iPhone, first iPad and first Apple Watch prove this. Now the 2nd gen Studio is much better in quietness. The first gen Mac Pro for AS looks rushed as well.

Always go for 2nd or 3rd gen products

M3-series will give some solid gen3 headless desktop options; M3 & M3 Pro in the Mac mini, also M3 Max & M3 Ultra in the Mac Studio; all four SoCs likely to have first generation Apple silicon hardware ray-tracing...

M3-series should also see the 2nd generation of the ASi Mac Pro; featuring M3 Ultra & M3 Extreme SoCs...?

So if we go by the 2nd or 3rd gen thing; the M4-series or even the M5-series is where you want to be for 2nd or 3rd gen hardware ray-tracing...

At which point Nvidia will be on what...? their 5th or 6th generation of hardware ray-tracing...?

I listened to the latest episode of ATP, so that you don't have to

Doing God's work...

The main reason to listen to this show is John Siracusa's epic rants, with Marco Arment as the "straight man", and Casey Liss as topic moderator. Siracusa's toaster reviews are legendary. That being said, he's hitting "old man yells at cloud" territory.

Regarding the Mac, they spent about five minutes on the new Mac Studio, and are happy that the sound issue appears to be fixed. Siracusa then spent a half hour ranting about how much he hates the new Mac Pro and it's not what he wants. Marco Arment points out that Siracusa purchased a 2019 Mac Pro with the sole intention of using it as a gaming PC. (Siracusa has stated that he only plays Destiny in Boot Camp.) Speaking of games, Siracusa rants for another half hour about how much he hates Apple's new gaming efforts and that they're going to fail. Then, they spend two minutes on Sonoma, and Siracusa is happy that they fixed his two pet bugs. Finally, Apple execs say that they saw the "Believe" Mac Pro shirts from ATP.

I can only SMH or /rolleyes so much...!

But buying a 7,1 just to play Destiny in Boot Camp...?

I guess I NEED a 14,8 for World of Warcraft...?

bEcAuS i NeEd pCiE SlOtS FoR My cApTuRe/sTrEaMiNg cArD...!?! ;^p

...the M2 Mac Studio is approximately ~8 times quieter than the previous generation. That's a notable decrease in sound output at idle, therefore Apple definitely tinkered with the cooling solution.

That is quite the improvement...!

I think the issue was the fan motors in the M1 Max/Ultra Mac Studio not being exactly the spec Apple wanted, but being what was AVAILABLE; the new M2 Ma/Ultra variants probably have better motors, doubt the "quad blower" housing or impellers themselves were changed...?

Possibility Apple has also addressed the coil whine issue with the PSUs; hopefully just needed some component changes on the PCB, another possible "working with what was available at the time" thing...?

 

[theorist9]

oh no. Apple's first gen products always suck in longivity. The first iPhone, first iPad and first Apple Watch prove this. Now the 2nd gen Studio is much better in quietness. The first gen Mac Pro for AS looks rushed as well.

Always go for 2nd or 3rd gen products

I was thinking of the Macs specifically. I believe they got the first gens right on the AS Mini, AS Air, AS iMac, and AS 13"/14"/16" MBP, while they fell a bit short on the AS Studio. So that's 6/7 = 86%, which I think qualifies as 'usually but not always'. I'd say the jury's still out on the AS Mac Pro; they do have an issue with the internal SATA connection, but if it's, say, a fully fixable firmware issue, that doesn't count. I'll defer to others on non-Mac Apple products.

 

[theorist9]

Aside from the bellyaching, they do point out something that I forgot to mention. When comparing the sound levels between the M1 Mac Studio and M2 Mac Studio, Apple's stats claim that, while idle or web browsing, the M1 Mac Studio is 15dB, compared to the M2 Mac Studio at 6dB. Decibels are logarithmic, as seen on this chart, which means that the M2 Mac Studio is approximately ~8 times quieter than the previous generation. That's a notable decrease in sound output at idle, therefore Apple definitely tinkered with the cooling solution.

That's a good point about the decibels. Though note that, while a reduction from 15 dB to 6 dB means a reduction in acoustic power of 10^((15-6)/10) = 8 times, that translates into a reduction in perceived loudness of 2^((15 - 6)/10) = 2 times. I.e., it's 2x quieter.

Formulas for converting decibel changes to ratios of acoustic power and perceived loudness (surprisingly, I couldn't find them online, so I had to derive them--I think I got them right):

Ratio of acoustic power = 10^((dB2 - dB1)/10)
Ratio of perceived loudness = 2^((dB2 - dB1)/10)

The acoustic rules of thumb follow from these: Every 3 dB increase corresponds to a doubling of acoustic power (10^(3/10) = 2), while every 10 dB increase corresponds to a doubling in perceived loudness (2^(10/10) = 2).

 

[exoticspice1]

M3-series will give some solid gen3 headless desktop options; M3 & M3 Pro in the Mac mini, also M3 Max & M3 Ultra in the Mac Studio; all four SoCs likely to have first generation Apple silicon hardware ray-tracing...

M3-series should also see the 2nd generation of the ASi Mac Pro; featuring M3 Ultra & M3 Extreme SoCs...?

So if we go by the 2nd or 3rd gen thing; the M4-series or even the M5-series is where you want to be for 2nd or 3rd gen hardware ray-tracing...

At which point Nvidia will be on what...? their 5th or 6th generation of hardware ray-tracing...?

yep. If you want good RT from Apple go for their 2nd gen RT cores, so M4. I say this because A12 NPU was serveral times better than the A11 NPU. The M3 will do a good job at introducing RT to the Mac world just like how Nvidia's 2000 series did with their 3000 series being much better at RT.

I know that Apple will likely surpass AMD in RT but nividia will be hard to surpass in terms of pure hardware but who knows.

One thing is for sure, hardware based RT will be very important for Apple and their Vision Pro. If they want their M chips to be powerful in GFX and creating lifelike games they better invest in the GPU department.

 

[B01L]

M∞ on 0nm process, that's the one to jump in on...! ;^p

 

[theorist9]

M∞ on 0nm process, that's the one to jump in on...! ;^p

I'm waiting for M

on i nm.

 

[Colstan]

Max over at his eponymous channel has put out a video comparing the M2 Ultra Mac Studio and M2 Ultra Mac Pro. Note that the Mac Studio is the 60-core GPU variant, while the Mac Pro features the upgraded 76-core GPU.



In short, the Mac Pro slightly outperforms the Mac Studio in CPU benchmarks, which they found to be consistently reproducible. However, the performance difference is slight and not something that would be noticeable in daily use. The Mac Studio uses more power, but both machines are efficient. The Mac Pro does run hotter, but again, the functional difference is small. Regarding the cooling solutions, they are both silent machines and he couldn't get the fans to ramp no matter how hard they pushed these new Macs.

Naturally, Max points out that the Mac Pro's design is inherited from the Intel era, which featured Xeons and multiple graphics cards. He concludes that, unless you need slots inside your Mac, there's no reason to purchase it over the Mac Studio.

 

[Yoused]

I'm waiting for M

on i nm.

The "M" stands for "mandelbrot" ?

 

[theorist9]

The "M" stands for "mandelbrot" ?

...hadn't even thought of that. Though I do vaguely recall meeting him at a conference many years ago. I suppose if I were going to reference a relevant mathematican (in my highly irrelevant post), I would have used C (for Cantor).

 

[amonduin]

I listened to the latest episode of ATP, so that you don't have to. The main reason to listen to this show is John Siracusa's epic rants, with Marco Arment as the "straight man", and Casey Liss as topic moderator. Siracusa's toaster reviews are legendary. That being said, he's hitting "old man yells at cloud" territory.

Regarding the Mac, they spent about five minutes on the new Mac Studio, and are happy that the sound issue appears to be fixed. Siracusa then spent a half hour ranting about how much he hates the new Mac Pro and it's not what he wants. Marco Arment points out that Siracusa purchased a 2019 Mac Pro with the sole intention of using it as a gaming PC. (Siracusa has stated that he only plays Destiny in Boot Camp.) Speaking of games, Siracusa rants for another half hour about how much he hates Apple's new gaming efforts and that they're going to fail. Then, they spend two minutes on Sonoma, and Siracusa is happy that they fixed his two pet bugs. Finally, Apple execs say that they saw the "Believe" Mac Pro shirts from ATP.

Aside from the bellyaching, they do point out something that I forgot to mention. When comparing the sound levels between the M1 Mac Studio and M2 Mac Studio, Apple's stats claim that, while idle or web browsing, the M1 Mac Studio is 15dB, compared to the M2 Mac Studio at 6dB. Decibels are logarithmic, as seen on this chart, which means that the M2 Mac Studio is approximately ~8 times quieter than the previous generation. That's a notable decrease in sound output at idle, therefore Apple definitely tinkered with the cooling solution.

I listened to it as well and generally found his complaints about the GPU power of the M1 and M2 to be rather disappointing. I know the M1 and M2 aren't going to beat an RTX 3060 but they don't have to in order to offer great Mac gaming performance. I have the My mid 2019 MBP has the best GPU you could get, the Vega 20, and the M1 is faster than it at gaming. I can play Overwatch and Valorant in bootcamp, Dota2, LoL, Starcraft 2, and more without issue. These games are generally pretty popular and are trivial for the hardware in the M1, the idea that the hardware is the problem with gaming on the Mac is kind of laughable.

 

[Cmaier]

I listened to it as well and generally found his complaints about the GPU power of the M1 and M2 to be rather disappointing. I know the M1 and M2 aren't going to beat an RTX 3060 but they don't have to in order to offer great Mac gaming performance. I have the My mid 2019 MBP has the best GPU you could get, the Vega 20, and the M1 is faster than it at gaming. I can play Overwatch and Valorant in bootcamp, Dota2, LoL, Starcraft 2, and more without issue. These games are generally pretty popular and are trivial for the hardware in the M1, the idea that the hardware is the problem with gaming on the Mac is kind of laughable.

Welcome to the site!

 

[Yoused]

I suppose if I were going to reference a relevant mathematican (in my highly irrelevant post), I would have used C (for Cantor).

actually, I choose to believe that the "M" is for Margaret

 

[Jimmyjames]

...the idea that the hardware is the problem with gaming on the Mac is kind of laughable.

The idea is a phantom. No matter what problem is refuted something else is found. Hardware, software, dev relations, shadows, “jaggies” (one for the banned MR members).

That’s not to say there aren’t challenges or that Apple has been a good steward of games, but I do think it’s reasonable to say things appear to be changing.

 

[Colstan]

iFixit finished their teardown of the M2 generation Mac Studio. It's identical to the previous model, other than the chip inside. The only change to the cooling solution is that Apple adjusted the fan curve, which could be done in firmware.

 

[dada_dave]

iFixit finished their teardown of the M2 generation Mac Studio. It's identical to the previous model, other than the chip inside. The only change to the cooling solution is that Apple adjusted the fan curve, which could be done in firmware.

Isn’t there a piece of software that lets you do that? I seem to remember one being discussed but I can’t remember if it worked for Apple Silicon or just reported the numbers.

 

[Colstan]

Isn’t there a piece of software that lets you do that? I seem to remember one being discussed but I can’t remember if it worked for Apple Silicon or just reported the numbers.

Macs Fan Control. Most of the features are free, with profiles being the paid "Pro" option. I've tried the free version, and it works as expected. They haven't added support for the M2 Mac Studio or Mac Pro, as of yet. There's also TG Pro, but it only has a free trial.

 

[quarkysg]

Not a good idea IMHO to lower the fan curve tho just to lower noise. I would think it is safe to do it the other way, i.e. turn it up.

 

[B01L]

That ending though...!

Also, an amazingly engineered chassis & subsystems...!

 

[Colstan]

Max Tech compared the new M2 Ultra Mac Pro against the old Intel model.



Unsurprisingly, the new model makes the previous one look like an old, noisy beast. Except for a slight uptick during a Cinebench torture test, the Apple Silicon Mac Pro didn't ramp the fans and was functionally silent. Comparatively, if you've ever been anywhere within the proximity of an Intel Mac Pro, then anything mildly stressful turns it into a jet engine. Obviously, the new model should perform better than the previous one, but it does so silently, which isn't something the Intel era was known for. My puny 2018 Mac mini with a 4-core i3 makes more noise than any of the Apple Silicon Macs. For most users, the Mac Studio is probably a better option, but if you need slots, then the new Mac Pro should fill that niche.