Lets talk about RISC-V boys
This is going to be one hell of a ride, a fully open source ISA, not the first time of course, but most the most significant with market ready prototypes
because openarm was a successful thing too.
>>58354226
Looks really cool but it probably won't get any traction outside of free software foundation
>>58354261
It could make the basis for a fully open RPi competitor.
>>58354283
yea no one will invest a dime on that.
>>58354352
With no licensing costs and someone else to do the drivers, a Chinese company might pick it up.
>>58354243
This is much bigger than OpenARM though, it even has a dev board out right now. OpenARM was limited in a variety of ways
>>58354261
I wish it does though
How much would it cost to fab them on 28nm?
We need to make a 64bit RISC-V cpu that has no concern for power use and is 100% focused on the highest IPC possible. Pretty much all new CPU designs trade performance for being able to run in laptops and servers. There will be nothing left for desktop use from major vendors if the trend continues.
>>58354410
>How much would it cost to fab them on 28nm?
>Scott gave me some ballpark numbers on the non-recurring engineering (NRE) costs associated with a new device. At 28 nm, the design cost is $50 million to $90 million. Each mask set is $2 million to $3 million. If you're lucky, you'll get away with going through the mask creation process twice to get the design done. Not lucky? Then more iteration.
http://www.planetanalog.com/author.asp?section_id=385&doc_id=561004
>>58354410
What would you be running on riscv that would need higher than x86 IPC I'm pretty sure riskv is supposed to low power by design
>>58354352
They're already making "Arduinos" with it, rPi's arent that big of a stretch. What matters more is for all the software to get ported to RISC-V.
>>58354425
>design cost
I though this was open, whats up with that?
>>58354453
Its open source right and Linux already has riskv support so wouldn't you just have to recompile software
https://www.youtube.com/watch?v=gg1lISJfJI0
THANK YOU BASED NVIDIA
>>58354495
Linux has RISC-V support already?
>>58354226
what is risc-v? Something like vt-x set of instructions for x86?
>>58354454
It IS open, so ignore the design cost
>>58354521
Not in mainline yet but it does run on risc-v.
>>58354551
Virtualization for ARM - when?
>>58354534
RISC-V is merely a specification. Some of the implementations of the design are and will be open, some will be closed and proprietary. Which one you choose to use depends on your application and the costs of implementations vary wildly depending on your objectives.
>>58354587
i meant "some implementations of the specification"
>>58354587
True
It will be like ARM but without royalties that you will have to pay
>>58354226
>x86_64 - 3 compatible competitors
>PowerPC - 1 competitor
>general-purpose MIPS 6400 - 3 incompatible competitors
>ARM - over 9000 incompatible competitors
>RISC-V - over 9000 incompatible competitors
We are coming to an age of micro-architecture hell, /lad/ies and /g/entoolmen.
>>58354410
Typically you'd use OpenPOWER for that.
>>58354669
I don't know why IBM don't produce a desktop motherbord to promote their cpu arch
>>58354604
You wouldn't have to pay royalties to RISC-V but you could still choose to pay royalties to someone like ARM for one of their RISC-V cores.
>>58355762
Probably due to a lack of demand.
>>58355762
Wasn't that basically what PowerPC was?
>>58355862
and the high costs associated with producing hardware.
>>58354604
>but without royalties that you will have to pay
That and the fact that you are, in fact, free to make modifications without seeking approval.