I want a piece of this.
Is RISC-V stock available?
How is it listed, and what is its price?
SiFive reckons its fastest RISC-V processor core yet is closing the gap on being a mainstream computing alternative to x86 and Arm. The yet-unnamed high-performance design is within reach of Intel's Rocket Lake family, introduced in March, and Arm's Cortex-A78 design, announced last year, in terms of single-core performance, …
RISC-V is an open standard for the instruction set. Anyone can design a chip that uses it. This is like how both Intel and AMD make chips that use the x86 instruction set. Most programs for x86 will run on any sufficiently-powerful Intel or AMD chip, and most RISC-V programs will run on any sufficiently-powerful RISC-V processor from any manufacturer.
You can't invest in "x86", but you could invest in Intel or AMD.
Similarly you can't invest in "RISC-V" itself, but you can invest in the companies who design the chips. One of whom is SiFive. No idea if they have publicly-traded stock, though.
(Standard disclaimer: This is not investment advice. If you invest your money, you may lose all that money).
As much as RISC-V is marketed as such, just having an ISA open source doesn't really bring scalable benefits.
It is a misunderstanding to compare RISC-V to x86 or x64, RISCV is far more fragmented.
I think it is already upto 9 letters now.
There is no standardised microarchitecture behind the ISA.
A program can work great on one RISC-V CPU and not at all on another.
The mix and match style of the ISA means you will need to release multiple versions of your SW (if you can get access to the toolchain and HW).
What will happen I think is some incumbent/de jure vendor will emerge, likely SiFive, and the rest of RISC-V will be used inside custom hw (where no one else can compile SW for the target)
The end result is that RISC-V is open source on paper, but the version used elsewhere is SiFive's version.
If I am honest I dont really see the point outside of custom HW.
RISC-V, if it takes up, is open source being used to accelerate vendor lock-in. We'll have routers using some custom RISC-V processors that cannot be compiled for once the vendor chooses to stop support. It can be a clever tool for planned obsolescence, to get around the GPL license I think, as the wider community have the source code, but not the toolchain for that proprietary HW.
An ARM A78 based Android phone is easy to port to by anyone, the vendor is ARM. A RISC-V one will need that specific RISC-V vendor's toolchain, and manuals, and microarchitecture.
If they wanted RISC-V to benefit the wider community, they should have locked the ISA, as every other ISA does today. Otherwise what actually exists is just a bunch of proprietary ISAs being given a common parent class name called riscv.
But the HW "object" I actually get is from a different class.
"Having it tested and verified on silicon, and a tool chain to go with support, for people call up and say 'how did you do this?' – that is the inherent value of an IP company like SiFive," he said.
Amen to that ..... and such a value is priceless/far beyond any standard measure of cost. However, as is always natural in such Elevating and Enlightening Instances of Advanced IntelAIgent Design, if you have to ask to know, one be not quite ready to enable and be able to deal with what one is able to enable and deliver for others to deal with in engagement, employment and enjoyment of the IP Provided ..... which is a Failsafe Security SafeGuard Default Supplied to Protect Systems Administrations from Inflicting Upon Themselves Catastrophic Terminal Self Harm.
And how nice it is to read that there is so much interest in the field, with Registered news today advising on at least three alternate pioneers in the genre [ARM/Alibaba/SiFive] with market wares to offer.
That was almost understandable. ..... Def
And there we was, thinking it to be certainly not misunderstandable, Def.
So .... in order to accommodate that anomaly ...... still more of an alien work in rapid progress needed?
Ok, so be it .... your wish our pleasure to command and control with virtually remote command and control, although rather more AWEsomely than Mechanical Turk-like can reveal its source code roots for testing with regard to identifying core provenance, should such an enlightenment be decided to be made more readily available .... and necessarily attractive to be able and enabled to be enthusiastically supported and reinforced/strengthened.
While it's fun to see a company grandstanding on its upcoming products, we've seen too much of this from large companies (remember Itanium? AMD's pre-Zen era?) to put blind faith into what some upstart company claims.
Faster than existing commercial designs or close to them from competitors is a bold claim to make. Unless there's some hidden talent working on the underlying (proprietary) design that hasn't been poached yet by Intel, AMD, Apple, NVidia or any of a dozen others, my bet would be on it being an 'okay' design, that'll happily compete with the desktop x86 & RISC-V CPUs coming out of China today, but not an Intel, AMD or Arm killer.
Happy to welcome more competition, of course.
This is more about media PR, SiFive is using "opensource" credentials to make it the dejure vendor. That is the strategy.
To become dejure requires them to be the incumbent processor associared with the term RISCV. They will need to build critical mass, to use sifive based toolchain, SW etc. For eg when Ubunutu make a release, they want it to be SiFive based.
After that whatever the RISCV foundation does becomes irrelevant.
It is a long way off yet to being a drop in replacement for *current* ARM ISA based designs (made by ARM vendor), x64 ISA designs (made by Intel, AMD vendors).
I do not fully appreciate the long term benefit wrt ARM - if the appeal of an open source ISA is modification, and a vendor goes rogue with the published riscv isa, what's the point?
Feels like open source ISAs are a gimmick. like "cloud ISA"