For me, it’d be a close one between the Steam Deck and the Pinephone Pro. Although, the thought of having a phone that runs GNOME sounds like a whole lot of fun, so I think that I’d have to give it to the Pinephone Pro
For me, it’d be a close one between the Steam Deck and the Pinephone Pro. Although, the thought of having a phone that runs GNOME sounds like a whole lot of fun, so I think that I’d have to give it to the Pinephone Pro
deleted by creator
Well… obviously some people simply prefer the open source philosophy - and we are seeing open source processors coming out of the open source instruction set : By backing RISC-V, they are sending a message to OEMs.
Nvidia’s argument for using RV was that they had the freedom to do things that they weren’t allowed to with ARM, and as a result they designed better microcontrollers. Surely that turns into the consumer’s advantage?
I’m just a couch potato, but as I understand it, RV can beat ARM on PPA because of its modularity : whether that turns into a significant advantage for the consumer, I guess depends on what the processor is being used for.
RnD could be sped up either because of “standing on the shoulders of giants” or because of not having to wrangle licenses : The consumer potentially gets the next generation tech quicker…
And production might be much cheaper, and those cost reductions could be passed on to the consumer.
deleted by creator
Yes - Arm has had to ease up both on licensing fees to smaller companies, and also has granted more flexibility in design, both as a result of RISC-V’s popularity. And they certainly didn’t want to do the latter!
The really interesting question is : Will RISC-V have gained enough momentum by the time the Nvidia/Arm deal has been decided upon (and probably rejected)? (And was this all a ruse by Nvidia for RISC-V??)
The recently announced extension ratifications help - especially Vector. And China is certainly pushing it, thanks to the tech trade dispute.
But here’s the point : There’s no real reason for designers not to use RISC-V. They’ve been teaching RISC-V at top universities for a decade. There’s open source software available to design and ratify. There’s no licencing costs, and there’s even funding available to get designs fabbed.
Not sure, but if Nvidia can just use RV (R5?) now instead of ARM, why waste money to buy ARM then? The advantage to them I guess would be to form a monopoly, so the acquisition will surely not be allowed.
If it were a ruse I don’t see how such a ruse benefits RISC-V… perhaps fear of Nvidia owning ARM will drive others towards RISC-V, but that benefits Nvidia how? Better free designs?
Well… if you’re proposing to spend $40B, you’ve probably gamed it out!
They’re allowed to buy Arm : they get full control of the ARM ISA, and can do with it what they like, regardless of anyone else using it… And yes, they get to find out competitors’ secret developments. They’d control what is being touted by some as the dominant ISA going forward. WIN. (I doubt it will be allowed to happen).
But for acquisitions of this size, there’s always a long approval period : a time of indecision and uncertainty (which we’re experiencing atm). Everyone in the industry is putting in a contingency plan, in case they do get control of the ISA… Everyone is putting resources into RISC-V atm. Joining the foundation. Advertising for designers with RISC-V experience. So RISC-V benefits enormously, and much progress has been made, which Nvidia (as well as everyone else) can benefit from. WIN. And saved $40B to boot!
Good analysis
:) Thanks
I just think it’s neat to have a new architecture, and the fact that it would compete with ARM is also exciting. However, you are right that there are potential pitfalls. I am waiting to see how this goes, but I’m ready to embrace an open arch over more closed alternatives, even if it’s not perfect.
I’m curious to see if the Mill architecture will take off any time soon. Unfortunately, it’s not an open platform, but it would represent some interesting innovations in processor design. They’ve apparently been making steady progress, but publicly facing information about progress is limited.
First time I’ve even heard of this
It’s been in progress for a while, but it’s not gotten to market yet. The team’s pretty small and there is a lot of work to do as with any new architecture. It does provide some interesting potential to vastly reduce the power usage compared to out of order architectures. It also provides various other interesting innovations like an alternative to registers, L1-only memory, a single address space, and an intermediate instruction set that gets specialized to the individual processor.