No.2173
>>2172The ability to to actually own the things you use. To not have government back doors thanks to the compliance of Intel and AMD. To seen reverb ideas and new methods of doing tasks by anyone being able to pick up the project and just run with it.
No.2182
>>2173How does RISC-V help with that? It's an open architecture, not an actual implementation and definitively not a build-your-own-CPU-at-home kit.
No.2188
>>2182but average human may have better chance of homebrewing tinfoil cpu out of RISC-V specification
and readily available open implementations than say x86 clone right?
No.2192
>>2182It doesn't have to be a ready to order CPU to make a difference. RISC-V is building steam, look at the semiconductor company SiFive. Once RISC breaks through in China I'm confident we'll start seeing more in mainstream devices.
No.2193
Speaking of China this article highlights the hurdles that RISC has to jump through to make it big in the Chinese assembly lines and R&D.
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333976
First, most Chinese chip vendors are already invested heavily in Arm and beholden to it. It’s hard to break free from Arm’s stranglehold.
Given Arm’s well-established ecosystem and the fact that it provides the preferred processing core for Android smartphones, why should local companies who design and develop chips for Android phones switch to anything else?
Second, RISC-V isn’t a free core. Yes, RISC-V offers a free and open-source instruction set. But this doesn’t mean that there are a ton of “free cores” lying around if a Chinese company knocks on the door at the RISC-V Foundation.
Third, developing a CPU requires design expertise in several specialties. These include electronic logic, compilers, and operating systems. Chinese designers need first-hand experience with modern, high-quality general-purpose computer instruction sets.
Fourth, RISC-V is designed to create a unified community. In a lot of community efforts, Asian engineers are often more eager to “take” from the community than “contributing” to it. As Wayne Dai, CEO of VeriSilicon observed, “Chinese engineers aren’t used to ‘uploading’ their ideas. They should start from there.”
Fifth, China is a master of fragmentation.
Time and time again, when a standard is set elsewhere in the world, eager Chinese companies — hopeful of carving out their own niche — have created a variant “standard” with a Chinese flavor. Although “derivative” China standards have rarely caught on anywhere, the value of having a uniform foundation to which everyone agrees has yet to crack the great wall of Chinese exceptionalism.
No.2285
>2282
I mean it is cool and options are always nice but MIPS is opensource now because of how horribly it failed during its "come back".
No.2287
>>2285Did it fail because of technical problems or business?
No.2323
the x86 patents are expired, and they won't protect an instruction set (which is all it is now…).
Why not make unlicensed clones? Don't speak to me about "well they added extra registers for SSE2 etc". Obvious extention.
Then send in demonic zaibatsu enforcers if anyone trys to sue to stop you?????
WHY DON'T YOU DO THAT?
No.2330
>>2323 Because x86 is a soykaf. You can run filthy chinese knock offs of a depreciated arch if you like though. I could not possibly think of you any less.
No.2757
>>2188>but average human may have better chance of homebrewing tinfoil cpu out of RISC-V specification and readily available open implementations than say x86 clone right?And it will be more feasible. Because developing CorpCorps New CPU or Lain Workatation using an existing ISA means you can taken advatage of existing toolchains, linux , etc far easier than desigining your own assembler, compilier, porting or writing an os plus software.
Instead just implement hardware and get the software for free.
