Intel seems to have a few financial issues

[jamcat]

Anyway. ARM CPUs are way overrated, IMO, and we will see that in tge future. Power savings is not something to worry for people demanding the best performance out of their machines - performance is. And, performance wise, ARM processor are no competitioin for the most powerful Intel CPUs.

This is what I mean by “oblivious and unprepared.”

The x86 architecture is approaching its physical limit. It won’t be able to get much smaller because of the size of the atom. And unlike ARM, the x86 architecture has inherently poor performance-per-watt. So the only way to continue increasing processing power will be to increase energy consumption. But that creates heat, and there is also an absolute ceiling on the operational temperature of the chip. Once again, physics.

There is a reason the industry is jumping ship. x86 had a good run, but it is approaching an iceberg.

[jupiter8]

Just watched Jeff geerling try the new Ampere Ultra 192 core ARM CPU. Compared to the AMD EPYC 192 core, the Epyc was way faster AND more efficient. The Ampere Ultra is way cheaper though so it beats the Epyc in performance per dollar but not in performance per watt, so there's that.

Just saying.

[uOpt]

And what software runs in Windows ARM? You realize that there isn't even an emulator like Rosetta 2.0. If it wasn't for Rosetta 2.0, Apple would have failed miserably.

There is. The question is how well Microsoft's translator works.

Rosetta 2 is a true hamsterpiece of software engineering, and Apple has soaked up many of the people who can pull this off.

Microsoft also has less pressure on them to make this work well. Apple was converting the fleet to ARM, MS has a very minor portion.

[uOpt]

Don't bet on ARM too hard. It is still licensed.

RISC-V will wipe it, sooner or later, depending on what jerk moves are being pulled with ARM licensing.

[Synthman2000]

Anyway. ARM CPUs are way overrated, IMO, and we will see that in tge future. Power savings is not something to worry for people demanding the best performance out of their machines - performance is. And, performance wise, ARM processor are no competitioin for the most powerful Intel CPUs.

This is what I mean by “oblivious and unprepared.”

The x86 architecture is approaching its physical limit. It won’t be able to get much smaller because of the size of the atom. And unlike ARM, the x86 architecture has inherently poor performance-per-watt. So the only way to continue increasing processing power will be to increase energy consumption. But that creates heat, and there is also an absolute ceiling on the operational temperature of the chip. Once again, physics.

There is a reason the industry is jumping ship. x86 had a good run, but it is approaching an iceberg.

I am not sure we need to sacrifice road tested reliability for more speed, do we ?

Probably audio related tasks are not the ultimate in speed craving.

I know speed is the done thing in the tech world and it is unlikely to pull back on faster CPU's. Personally I would use tried, trusted and reliable over pure speed every time.

I always find that the big new thing is hyped to extreme levels and then when you gte it it it is like, oh, ok then, meh difference. And what may come with new CPU designs is plenty of headaches.

So how fast would it need to be in order to counter the hassle factor is something I will ask, especially given critical mass for power heavy music work is a very long way off.

[vstdls]

Anyway. ARM CPUs are way overrated, IMO, and we will see that in tge future. Power savings is not something to worry for people demanding the best performance out of their machines - performance is. And, performance wise, ARM processor are no competitioin for the most powerful Intel CPUs.

There is a reason the industry is jumping ship. x86 had a good run, but it is approaching an iceberg.

I know speed is the done thing in the tech world and it is unlikely to pull back on faster CPU's. Personally I would use tried, trusted and reliable over pure speed every time.

I always find that the big new thing is hyped to extreme levels.

Yes.

Analogous to your point: 2023 (US) saw 200K more Ford F150s sold than all of Teslas vehicles put together, and only 100K unit s under the entire US EV market (that includes hybrids). Ain't nobody talking about Ford F150s. To say that the F150 is doomed would fit well into alot of the hyperbole in this thread.

(x86 is the pickup)

[PAK]

Anyway. ARM CPUs are way overrated, IMO, and we will see that in tge future. Power savings is not something to worry for people demanding the best performance out of their machines - performance is. And, performance wise, ARM processor are no competitioin for the most powerful Intel CPUs.

How are you coming to that opinion though? EG Would you like to name a CPU heavy synth and patch, and state how much polyphony / instances your current Intel / AMD CPU can do? I’ll take time to respond with Apple Silicon numbers. You get to define the parameters. My assumption is you’re forming impressions based on things that may not actually apply for your own DAW use..

Don't let Apple marketing BS full you.

You’re right people shouldn’t believe marketing numbers, or anything from those who present numbers as objective whilst they make money from certain outcomes. Often such numbers aren’t “technically” lying, but instead lie through omission or the nature of the comparison. Remember that it isn’t only Apple who engages in this.

Even in synthetic benchmarks, M4 is out-performing X86/64 in some single core tests. The reason you’re not seeing this replicated, in multi-core, is higher core count M4’s have still to be released. You’ll start seeing a change-over in consumer / high-end desktop numbers next year, even in synthetic benchmarks. M4 Max numbers are already known, so this isn’t guessing. It’ll be interesting to see some of the reactions when this happens, though you shouldn’t expect to hear much about it from those who sell PC’s

Very high core count, low clock rate, processors (IE some Threadripper and Xeon) will still dominate multi-core results. These aren’t ideal for audio usage given there are single-core CPU heavy synths which might exceed the limits of those lower-clocked cores. Plus they’re in a different price category anyway (Which Apple might compete against with so-called “Extreme” versions of their processors..)

The point being, you're genuinely wrong about ARM, particularly for things like DAW usage. However, Nvidia is hammering everyone in graphics/AI performance and, if those are more important to someone, the X86/64 path often makes more sense presently. Intel will survive regardless, though it's role in the desktop / laptop market may change. Microsoft are trying to push WOA (particularly on laptops) for a reason. They don't want to see what's happened to the US mobile market (where iOS is in the majority) happen to the US laptop market..

[Synthman2000]

ARM RISC may have a short service life if this has legs:

"Google's new quantum chip has solved a problem the best supercomputer would have taken a quadrillion times the age of the universe to crack"

I had to double check if it is April fools day.

https://www.livescience.com/technology/ ... e-to-crack

No more CPU overloads, ever.

[elxsound]

ARM RISC may have a short service life if this has legs:

"Google's new quantum chip has solved a problem the best supercomputer would have taken a quadrillion times the age of the universe to crack"

I had to double check if it is April fools day.

https://www.livescience.com/technology/ ... e-to-crack

No more CPU overloads, ever.

Not quite. Consider this an R&D achievement.

It’s nothing coming to the consumer market, yet.

[Bunny_boy]

ARM RISC may have a short service life if this has legs:

"Google's new quantum chip has solved a problem the best supercomputer would have taken a quadrillion times the age of the universe to crack"

A dad joke that is funny?

[pekbro]

You guys know ARM is owned by someone, and they are the developers. You can lic ARM from them, like apple has...

Anyway, the main driving force for ARM was always cost, not performance. If it's a performance race
now, then that's a recent thing...

[Synthman2000]

ARM RISC may have a short service life if this has legs:

"Google's new quantum chip has solved a problem the best supercomputer would have taken a quadrillion times the age of the universe to crack"

I had to double check if it is April fools day.

https://www.livescience.com/technology/ ... e-to-crack

No more CPU overloads, ever.

Not quite. Consider this an R&D achievement.

It’s nothing coming to the consumer market, yet.

There are rumours it can support 2 Diva's playing chords.

[jamcat]

Yall are like “Macs are too expensive” but you’re gonna run out and buy a quantum computer.

[Synthman2000]

Just a report from what I heard on the radio.

What does that Quantum chip even mean ? Let's hope they put that processor to use to end world wars, global warming and cancer. Otherwise what is the point, more rich blokes ?

Maybe we should question more deeply what is actual intelligence ?

[jamcat]

Yes, it will end war and cancer. But you don’t want to know how.