Fallacies about RISC-V
eetimes.com
eetimes.com
But this article is disingenuous to the point of willful misrepresentation.
It asserts that the only business risk to consider when choosing an ISA is whether the owner of the ISA might be sold to a new owner. Which I think might not even make the top 10 risks. Certainly it’s far less important than whether the ISA will reach critical mass for toolchain support, whether it will scale from small devices to datacenter applications, whether it will evolve and innovate, whether there is a pool of programmers with expertise, and so on. It’s bizarre to not even mention risks in these areas.
I’m not saying RISC-V falls short in these things; I don’t know enough to have an opinion on that. But I do know they’re all more important than whether e.g. a hedge fund takes Intel private.
Lots of other similar errors and omissions. This is evangelism from a True Believer, not a legit piece about RISC-V’s place in the industry.
> whether it will evolve and innovate
CPU evolution and innovation is driven by the transistors beneath the ISA, not the ISA. For Intel, MMX/SEE and all the SIMD stuff was an innovation and something new, but the instructions that use it are really not that big of a deal in and of themselves. When some other capability arises in CPUs, any ISA will simply get new instructions.
I assume you mean Linux-capable hardware.
The Risc-V foundation claims that over 10 Billion chips have already been produced.
The presented plan included quality choices like "vendor toolchains forked from upstream that support their secret sauce features, but suck at everything else compared to mainline, and are only maintained until until the vendor runs out of VC money," "You can always use the nearest base profile and ignore the special hardware you theoretically paid this vendor a fortune for" and special attention to "Most of your workloads are torch or something anyway, just use the vendor's binaries."
I was not impressed.
Extinct is a bit harsh. SPARC V8 is alive and kicking in the space industry as the LEON processor models. Not the most performant CPU, but good enough for a lot of applications.
And David is quite right that the Digital architectures are really extinct.
[Now I wonder if Compaq would open source the IP of Alpha. I think they licensed a lot of it to Intel so sadly they probably still make a bit of money from patents there.]
All of the ISA specifications are freely available[1] and you can buy a license to manufacture SPARC hardware for $99.
also though any us patent filed since 01995 expires 20 years after filing, so any us αxp patents filed in the period 01995–02001 have expired already, and us patents filed before 01995 expire 17 years after issuance, so we only have to worry about pre-01995 us patents if first dec and then compaq and intel managed to delay their issuance, lemelson-style, until 02005 or later, which seems extremely unlikely
so αxp is almost certainly already open-source as an isa
dec's implementations are i suppose not open-source though, and casual glancing around on opencores and the like doesn't turn up any obvious candidates
For the time being they are still selling LEON but I'd expect it to be obsoleted at some point.
But it's not just that, it's the use of an intermediate language in general. When you use an intermediate language, you are less tied to one specific architecture, and are instead tied to the existence of a JIT, Recompiler, or Interpreter for that intermediate language being available on the target architecture.
This has to do with shipping an executable in native code vs shipping something that isn't directly executable, because it's in an intermediate language instead of native code.
I absolutely want RISC-V to succeed but there is a difference between wanting it to succeed and wanting it to become 'dominant' or a 'lingua franca'. I find it a bit unnerving that to 'dominate' other ISAs is something that is being discussed.
The principle of open source yes, but not a particular implementation.
Also an ISA is quite different to a networking protocol where inter-operability is key.
The developers generally don't feed into the "flame wars" but it is a long tradition in the open source world to pick a piece of software and viciously attack the merits and users of competing software. This also happens with users of proprietary software, as with Apple cultists and so on. It is just human nature.
> I’m sure he’d be happy with any set of open archs dominating.
But he wasn’t. He was advocating for his own single arch dominating which is quite different.
Clearly having an open ISA is great news and one that we should support and hope it succeeds.
Advocating for that single ISA to crowd out all other ISAs and render them unviable isn't in my view for a number of reasons.
I still think RISC-V is superior to x86(-64) FWIW.
Or more properly to sell* ARM processors (or devices with ARM processors) you need a license from ARM where they give you permissions and (unless you are Apple) ARM is not going to give you a license where custom instructions are not forbidden.
RISC-V is open in the sense that you do not need permission from the IP owner to use it.
* I believe that you probably can manufacture custom ARM processors without a if you never plan to sell them.
That was the case if you are an ARM architecture licensee. And of late, that's not even the case anymore.
Happy to be corrected though if I've got this wrong.
The huge advantage of RISC-V is the lowered barrier to entry for chip manufacturers. It's now becoming possible for tech companies to make their own processors for their specific hardware and requirements.
In my mind, the even bigger advantage is that the door is now open for open source RISC-V implementations, so that people/companies can work together on high performance/low energy CPUs. It seems such a waste of talent for different companies all spending time and effort to implement similar features in their chips, when they could be building and extending open source chip designs.
For example whilst both AMD and Intel produce X86 CPUs and share the license they also have numerous patents for their specific implementations e.g. memory controller, prefetch unit, branch predictor etc.
The idea that the RISC-V ecosystem isn’t going to get rather quickly bogged down by patents is too optimistic unless I’ve missed something substantial in their licensing terms that prohibits RISC-V ISA users from issuing patents on their in silicon implementation.
Building a CPU today without having to work around patents is already rather hard and eventually there will be only so many actual hardware implementation of the RISC-V ISA that are both efficient in terms of performance and manufacturing.
So whilst the ISA might be open the secret sauce of how to actually make something useful out of it isn’t going to be open or free.
Whether or not RISC-V implementations become encumbered by patents, at least it is now possible for companies to make their own processors without having to create the entire ecosystem too.
What in particular do you think is incorrect in my comment and can you provide a correction please?