First Apple Silicon Benchmarks

[clockdogg]

am i really getting cancelled on here? be honest

No... mere minions like me have no power to cancel... only to tease and annoy. Besides, I like your posts - most of the time. You have a nose for tech news.

 

[Zarathustra[H]]

Not sure about right away, but remember... this is the company whose mobile CPUs have advanced so much that a $399 iPhone outperforms every Android phone. Apple has a kind of freight train mindset where it may be slow to start, but eventually becomes hard to stop.

Yeah, exactly. I think it is totally possible for Apple to be competitive in high end desktop CPU's eventually. I am just skeptical that they will do it in their first launch. Even if they have been at it for five years as aokman suggests, they have a deficit of 40+ in continuous improvement work done by Intel and AMD to catch up with.

With the amount of spare cash Apple has to spend, they can hire great engineers and get it done if they really want to and deem it worthwhile, I just doubt it will happen straight out of the gate.

We'll see though.

 

[IdiotInCharge]

Not sure about right away, but remember... this is the company whose mobile CPUs have advanced so much that a $399 iPhone outperforms every Android phone. Apple has a kind of freight train mindset where it may be slow to start, but eventually becomes hard to stop.

Outperforms at... what?

This is the repeated issue: compute benchmarks are only worth what difference they make to the user experience. I don't care about a Ryzen 3950X because it's not any faster than CPUs that cost half as much for gaming, and that's the most intensive thing I do. I don't care that Apple phones 'benchmark' better than my Pixel 3a, because they're not any better at being phones. They're just better at lots of things I don't do and don't need.

Where Apple will succeed or fail is whether or not they can make an appliance that better suits the needs of their users.

As skeptical as I am at the comparisons to desktop-class CPUs, I still think Apple is perhaps headed in a successful direction in the sense that for the limited usecases that their customers are accustomed to, they perform very well with their custom ARM cores.

What I keep reminding people about, though, is that Apple has a significant challenge ahead when it comes to usecases outside of their planned sandbox.

 

[OutOfPhase]

What I keep reminding people about, though, is that Apple has a significant challenge ahead when it comes to usecases outside of their planned sandbox.

They do, but I have to admit - I've been gobsmackered more than once by the competence of their chips in the last ~5 years.

That said, it is absolutely a wildcard how well those designs scale out of "amazing perf/watt" which matters so much in the markets they've been in to "now you're fighting Intel which is happy to throw 350 watts at you to win". And of course, real pros needing horsepower don't care about 350 watts, at all.

I think realistically, it will make seriously nice macbooks for most people using macbooks. Beyond that - I'd agree and say a much bigger longshot.

 

[sed8em]

Yeaaaah na lol, Apple have been working on this for probably 5yrs behind the scenes... they wouldn’t be replacing Intel unless they could beat them. People will be in for a shock when they hit the market.

I tend to agree with your sentiment.

I’m skeptical a multi billion dollar company hasn’t done their research and isn’t 100% ready to release a highly competitive product. I doubt this will be a flop.
I’m sure they’ve been working on this for years as well.

 

[Zarathustra[H]]

"Does Guy Fieri want to kick in your door and take you to flavortown while you sleep?!?"

This might just be my worst nightmare

 

[IdiotInCharge]

I’m skeptical a multi billion dollar company hasn’t done their research and isn’t 100% ready to release a highly competitive product.

So... you were born yesterday?

That said, it is absolutely a wildcard how well those designs scale out of "amazing perf/watt" which matters so much in the markets they've been in to "now you're fighting Intel which is happy to throw 350 watts at you to win". And of course, real pros needing horsepower don't care about 350 watts, at all.

For Macbooks, I expect that they'll have a fairly easy sell: perf/watt translates largely into battery life, and well, assuming they put in enough custom logic to cover their userbase and prevent heavy stuff from running as generic code on their ARM cores, I'm going to expect a few days at least. Right now I'd peg your average ultrabook to hit about 10 hours give or take, and I expect Apple to be able to triple or quadruple that simply by exerting control over the whole stack, like they've done with their phones and tablets.

For iMacs (etc.), I really don't know. ARM has not been shown to have the branching performance of x86 in real-world implementations. Really, nothing has, and while it's not hard to pull off impressive IPC with restricted workloads, it's the unrestricted workloads that define x86. Whether that remains relevant for their userbase (or for any of us) going forward is the magic question!

 

[Aurelius]

Outperforms at... what?

This is the repeated issue: compute benchmarks are only worth what difference they make to the user experience. I don't care about a Ryzen 3950X because it's not any faster than CPUs that cost half as much for gaming, and that's the most intensive thing I do. I don't care that Apple phones 'benchmark' better than my Pixel 3a, because they're not any better at being phones. They're just better at lots of things I don't do and don't need.

Where Apple will succeed or fail is whether or not they can make an appliance that better suits the needs of their users.

As skeptical as I am at the comparisons to desktop-class CPUs, I still think Apple is perhaps headed in a successful direction in the sense that for the limited usecases that their customers are accustomed to, they perform very well with their custom ARM cores.

What I keep reminding people about, though, is that Apple has a significant challenge ahead when it comes to usecases outside of their planned sandbox.

Yes, it's what the performance brings to the experience that matters... but that matters a lot.

If my phone feels fluid and responsive, that's important. Faster CPUs allow more sophisticated photography; they enable practical uses of emerging technology like AR; and if you're a mobile garmer... well, it means everything.

With Macs, I think goal is simple: expand that sandbox by offering superior hardware. Apple has always had to sell computers that were either behind the industry or just competitive; this is its chance to outperform x86 and make Macs the go-to computers for those who value speed in mainstream hardware, at least in certain cases.

 

[Deleted member 243478]

I tend to agree with your sentiment.

I’m skeptical a multi billion dollar company hasn’t done their research and isn’t 100% ready to release a highly competitive product. I doubt this will be a flop.
I’m sure they’ve been working on this for years as well.

There have been rumours of this since 2012, I think Apple have been working on this for a long time and probably realised it was going to happen 5yrs ago when Intel was dropping the ball.

T2 is basically an iPhone chip that runs in tandem to do all the security enclave stuff that Intel cant do and also does some heavy lifting for encoding HEVC etc. When T2 first showed up in Macs, we knew that Apple had already done a lot to integrate their chips into Mac products behind the scenes.