Apple Releases M1-Powered Apple Silicon Macs, macOS Big Sur Releasing This Week

[vladpetric]

Except that this particular chip will only achieve those performance per watt numbers on very select workloads that can use that chips dedicated hardware. In almost all workloads it just simply will not compete with x86 in overall performance. I've been watching various comparisons and for most people it really is just another ARM chipset. And even then it's still limited strictly to Apple's OS, which as we all know and is very well documented, performs really bad.

Well, do you actually have some quasi-official benchmark numbers here?

The main difference between regular ARMs (mostly toy, low-power-only designs) and Apple ARMs is the micro architecture, which allows it to have much much higher IPC.

The fact that ARM didn't produce higher performance chips in time is, IMO, the reasons it got sold around.

For a longer answer:

https://www.quora.com/Why-do-Apple-processors-tend-to-outperform-the-competition-in-performance-tests/answer/Samuel-Williamson-39?__filter__=all&__nsrc__=1&__sncid__=6319243134&__snid3__=9758384935

[duby229]

Well, do you actually have some quasi-official benchmark numbers here?

The main difference between regular ARMs (mostly toy, low-power-only designs) and Apple ARMs is the micro architecture, which allows it to have much much higher IPC.

The fact that ARM didn't produce higher performance chips in time is, IMO, the reasons it got sold around.

For a longer answer:

https://www.quora.com/Why-do-Apple-p...3__=9758384935

I mean even Qualcomm's chipset outperforms it on generic workloads. (Much of the difference is probably due to how horribly slow Apple's OS is when compared to Andriod) Look on youtube for video comparisons. It does well on workloads that can use its dedicated hardware. Like for example video encoding which has dedicated hardware or cryptography which also has dedicated hardware. But in generic workloads it really isn't that fast, it's just another ARM chipset. Most people won't even be able to use it at all, I mean Apple has less than 3% marketshare for a very good reason.

[birdie]

So already the hardware is not good for anything more demanding than developing iOS applications, but the lack of software choices would have also been enough to disqualify the product.

And MacOS X applications for that matter and that's exactly what these new laptops are designed for. Also, 16GB of RAM are more than enough to develop for web (JavaScript, Python, RoR, PHP) and code in Java, so they are not so bad.

What I'm really curious about is whether it will be possible to use virtual machines running x86-64 code on M1 and what the performance will be. That's what really important for developers who don't want to tie themselves to Apple's walled garden.

[vladpetric]

I mean even Qualcomm's chipset outperforms it on generic workloads. (Much of the difference is probably due to how horribly slow Apple's OS is when compared to Andriod) Look on youtube for video comparisons. It does well on workloads that can use its dedicated hardware. Like for example video encoding which has dedicated hardware or cryptography which also has dedicated hardware. But in generic workloads it really isn't that fast, it's just another ARM chipset. Most people won't even be able to use it at all, I mean Apple has less than 3% marketshare for a very good reason.

Do you have benchmarks for that?

Generally, this is something that should be settled with benchmarks.

[_r00t-]

https://en.wikipedia.org/wiki/Tegra#Orin I hope they'll make it earlier... Tegra Orin...

[duby229]

Do you have benchmarks for that?

Generally, this is something that should be settled with benchmarks.

As far as a thorough review with benches on popular sites like Anandtech or Techradar, all that exists are just vague claims by apple. You know the "3 times more powerful" claim, which is so vague it's meaningless. But in the meantime Apples new Iphone got released and comparisons of -that- are available. Check youtube.

Just keep your bias in check and realize that it has dedicated hardware for video encoding and that doesn't mean anything for CPU performance.

[vladpetric]

As far as a thorough review with benches on popular sites like Anandtech or Techradar, all that exists are just vague claims by apple. You know the "3 times more powerful" claim, which is so vague it's meaningless. But in the meantime Apples new Iphone got released and comparisons of -that- are available. Check youtube.

Just keep your bias in check and realize that it has dedicated hardware for video encoding and that doesn't mean anything for CPU performance.

Either you have a link to properly collected benchmarks or you don't. Otherwise your claim is weak.

And let them who are without bias cast the first stone.

[duby229]

Either you have a link to properly collected benchmarks or you don't. Otherwise your claim is weak.

And let them who are without bias cast the first stone.

You actually think Apple let people release benches before launch? In the meantime comparisons of the new Iphone -are- available. Check it out. I'd say that yes it competes well with Qualcomm's chipset, but it does -not- quite beat it.

And yes, I already cast that first stone.
Since you seem incapable of incredibly simple searches, here's one for you....

https://duckduckgo.com/?q=bionic+vs+snapdragon&t=h_&ia=web

[vladpetric]

You actually think Apple let people release benches before launch? In the meantime comparisons of the new Iphone -are- available. Check it out. I'd say that yes it competes well with Qualcomm's chipset, but it does -not- quite beat it.

And yes, I already cast that first stone.
Since you seem incapable of incredibly simple searches, here's one for you....
https://duckduckgo.com/?q=bionic+vs+...on&t=h_&ia=web

Typically the burden of proof is on the one making a claim. If you make a claim, you present the data you have, it's really as simple as that. "You look for that data for my claim" is a propaganda technique. Good job, comrade.

As for the first result in that list:

"So we can reach to a conclusion that Snapdragon 865 has reduced the performance gap to 25% with respect to A13’s performance. However, it must be noted that we are talking about the single-core performance of both the chips. If you count in multi-core performance then they are almost neck and neck."

(this is A13)

Actually 25% in single thread performance is pretty big.

[duby229]

Typically the burden of proof is on the one making a claim. If you make a claim, you present the data you have, it's really as simple as that. "You look for that data for my claim" is a propaganda technique. Good job, comrade.

As for the first result in that list:

"So we can reach to a conclusion that Snapdragon 865 has reduced the performance gap to 25% with respect to A13’s performance. However, it must be noted that we are talking about the single-core performance of both the chips. If you count in multi-core performance then they are almost neck and neck."

(this is A13)

Actually 25% in single thread performance is pretty big.

And yet most reviewers agree that in actual usage it's pretty close, with Qualcomm generally considered the better chipset. You already didn't believe me and wouldn't believe me anyway. Keep going don't stop at just the first result. Don't let your bias stop you from self education. The only way you're gonna learn the -actual- reality is if you look it up for yourself.

EDIT: We'll learn more about M1's -actual- performance after it gets launched, instead of Apples current "3 times more powerful" propaganda...