No word on NVIDIA’s N1X SoC supporting anything else than MS Windows unfortunately. Linux SoC ARM support is still rudimentary and NVIDIA’s track record is not promising.
Doesn’t SteamOS work on Arm?
I’d love to see an ARM CPU with unified memory outside of the apple ecosystem. Maybe in 5 years we’ll see a framework laptop that competes with the M chips on efficiency.
RISC-V!!
Sure, in 20 years when it catches up with other ecosystems
Seems like a niche use case (in context of laptops). If you really need unified memory, wouldn’t you be using a desktop?
I don’t see how that’s logical at all. Why restrict a technology to the desktop?
I don’t mean restrict, I am talking more about market demand.
Desktops are a bettee fit in terms of cooling and power availability for long throughout tasks (not burst type use cases).
These is a reason why both Framework and Nvidia have released desktops with large unified memory options (as well as many smaller mini-pc manufacturers) and not laptops.
Does the word “Macbook” ring a bell? They’ve been shipping unified memory in laptops with pretty resounding success for quite a while now.
Unified memory technology clearly doesn’t require desktop-level cooling and power availability, although it can of course make use of it.
Sure, but overwhelming majority of people buy Macbooks because of the design and ecosystem, not because of unified memory.
Of course it doesn’t require desktop-level cooling and power availability, but chances are if you’re not running MacOS (which has a broad set of use cases, including ones where unified memory is not the key adoption driver), you do want to make use of the additional power and cooling. That’s why I cited the Nvidia Spark, the Framework desktop and the AMD-based mini-PCs with unified memory). There is a reason why Framework, known for making laptops, went with a desktop solution for a unified memory desktop.
I am willing to bet, Nvidia is not going to offer large memory sizes on their N1 SoC, they want people to buy the Spark and that’s not their target area.
I don’t doubt that Nvidia is capable of a powerful chip solution, but price (especially with the current situation) and broader ecosystem support are open questions. This applies both to WoA and the higher costs available with servicing high end ARM laptops (much of the world is a lot more price sensitive than the US).
This being Nvidia, I don’t really expect much. I am looking forward to seeing performance and efficiency benchmarks.
