AMD claims up to 24 hours of laptop battery life with its latest Ryzen 6000 silicon AMD's next-generation Ryzen 6000 microprocessors, announced at CES this week, should provide another option for the Intel faithful to rethink options when buying laptops. The chip designer claims the new processors offer up to 24 hours of laptop battery life on video playback with a single charge, and double the graphics …
We aren't in the days of a 30,000 transistor ARM2 anymore - a modern ARM is pretty complicated and not very RISCy
Is there really that much difference in TDP between a 5nm TSMC x64 chip and a dozen core Apple ARM?
Or is the Apple M1 battery life mostly from smarter GPU use and better integration design eg shutting down peripherals?
Indeed, in general there's little difference overall in chip design: chips with the same capabilities build with the same processes will be of a similar size, though support for x86 in hardware will still mean some overhead. But ARM still has an advantage when it comes customisation and creating discrete components on the die for codecs, encryption, etc. Apple's major advantage for apparent improvement is the return to shared memory for CPU/GPU which makes some operations much faster because nothing needs copying. This is great when it works but might also mean that you need more RAM than you thought you would.
Quote: "Are there desktop machines with this chip?"
Short answer 'no'.
Assuming you're talking about the mobile 6000 series Zen 3+ based mobile chips, then no, those are not, as far as I know, coming to desktop. (A Zen 3+ refresh was rumoured quite some time ago, but from what I understand AMD scrapped the refresh in favour of Zen 4, but again this is just a rumour).
The current (as in been around a while now) desktop 5000 range is Zen 3 (the mobile 5000s are a mix of Zen 2 or 3 depending on model), and other than one enhancement to Zen 3 coming [*] the next desktop parts will be Zen 4 based, and use a new AM5 socket.
So if you are after a new desktop AMD, then it's either current (i.e. now slightly older) Zen 3 with the 5000s, or wait for the new Zen 4 based 7000 models, due some time later in the year, likely summer or later.
Worth noting, Zen 3 desktop parts use the AM4 socket and DDR4, whereas Zen 4 will use the new AM5 socket and DDR5, so there is no direct upgrade path from AM4 Zen 3, without getting a new motherboard and RAM (in addition to the CPU).
* The enhancement, is a new version of the existing Zen 3 based 5800X, called the 5800X3D (due this Spring), this is basically the same part as the 5800X (Zen 3 8 core AM4 7nm DDR4), but with a 64MB cache added. This is aimed at gamers, although I'd assume any tasks that are cache heavy would likely benefit.
AMD are touting the 5800X3D as the "Worlds fasting gaming processor", but take that with a pinch of salt till some independent and reputable benchmarking has been done with real silicon.
Is AMD feeding the power optimization bits to LKML and will they hand the discrete/integrated switching tools in as well?
I'm not running radeon on my desktop atm, but I've heard some good things about the opensource drivers for it. And I've heard that the proprietary drivers are slowly getting better on linux as well.
If all the above, I'd be quite happy to get one of these in a laptop.
That's difficult. Laptop energy usage can vary wildly depending on workload. If battery life claims were only made under worst-case scenarios, they would mostly be around one hour. That would be honest, but not very useful, because all kinds of optimizations that can have a large impact in realistic conditions wouldn't be tested at all.
IMO, the real problem is not that laptop makers are picking generous conditions for their tests; it's that they are not picking them in a consistant way, which makes comparisons impossible.
What we really need are standardized benchmarks, employed and respected across the industry. We actually have this for things such as CPU and GPU power, so it shouldn't be outright unfeasible. The benchmarks wouldn't need to represent a 100% workload; they ought to represent a realistic workload, or even better a small variety of realistic workloads (e.g. gaming, video streaming, office, browsing).
This could still be gamed, in some fashions, but at least we'd stop comparing apples to oranges.
"That's difficult. Laptop energy usage can vary wildly depending on workload."
That's why you run it flat out as a worst-case scenario. My philosophy is that it's a lot better to expect the worst and be pleasantly surprised than to be given good news and then be disappointed. Same here. Tell me the bare minimum I can expect out of this thing and then I can call all the extra time gravy.
"That would be honest, but not very useful, because all kinds of optimizations that can have a large impact in realistic conditions wouldn't be tested at all."
But as you yourself said, to each his own. Everyone's demands and expectations differ, so given such a scenario, I'd like to see a floor. Besides, it might convince people to demand better batteries (or more-efficient top-end tech) if they learn they can't run a laptop flat out for more than an hour.
"The benchmarks wouldn't need to represent a 100% workload; they ought to represent a realistic workload, or even better a small variety of realistic workloads (e.g. gaming, video streaming, office, browsing)."
A spectrum, then? That could be useful, but like I said, I'd still like the see a floor as part of those benchmarks.
A "security" processor designed and running code by Microshite with access to "critical data like system credentials, user identities, encryption keys and personal information" what the fcuk has happened to this reality?
You can't trust any company under US control, we are living in a dystopian novel.
AMD has their own “management engine” version included. If I recall the ratio correctly, on a chip with 16 cores there are 4 ARM cores running secretly in the background, having access to everything.
Don't buy Western shit. They are losing their grip on the population. Measures are becoming more and more desperate.
This post has been deleted by its author
while I support AMD in heart (though not in wallet, as I had to settle for an intel laptop for my dahling, due to shortage of amd chips), I call this 24 hr claim a blatant lie, and the usual givaway it's just a lie, is the 'up to' bit, with a list of exemptions longer than a (...). Until I see _real world_ charts, and _real world_ comparisons (including those on the claimed 'double' graphics performance) - I will not fall for any such bullshit, long past that stage.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
