SPARC, Intel i960, AMD29k, ARM, RISC-V and many others are direct descendants of the Berkeley RISC design.
Using the number of years of a measure of technological relevance, it can be argued that ARM and RISC-V designs are even older than 35 years. It is pure sophism, though, and it does not help.
By the same token, all modern computing architectures are fossilised designs because they are based on the von Neumann architecture from 1945.
Just want to say thank you. I could have typed that out and didn't. ( See my reply below )
My faith on HN discussions continues to fall. There are many topics I keep thinking may be I shouldn't even click on. But it is good to know there are people, who continue to contribute in a positive manner. ( at least more positive than I am )
> Using the number of years of a measure of technological relevance, it can be argued that ARM and RISC-V designs are even older than 35 years.
I don't know what you are talking about.
X86/ARM are highly successful, being 30-40 years old proves that they are well maintained & comes with a long history of good ecosystems. RISC-V is obviously rising with no limit for its future.
POWER lost the competition, it is done, it should be depreciated. With that in mind, its 35 years history is not a fancy record, it is proof that not many people are going to pick that junk up again.
x86 and ARM are well maintained because they're 30-40 years old, but Power (it's not an acronym anymore) is not because it's 35 years old?
I also suspect you're making the typical mistake of confusing PowerPC with modern Power ISA. They overlap strongly and PowerPC was virtually designed to be 64-bit ready from the very beginning, but chips based on the modern ISA are much more sophisticated.
As someone who actually works on ppc64, the biggest two things that bother me about the ISA are the weirdness with r0 and the large number of instructions that must be implemented to be practical on a new design. Those are comparable to the quirks of any other "successful" ISA by your standards. They hardly make it junk.
> x86 and ARM are well maintained because they're 30-40 years old, but Power (it's not an acronym anymore) is not because it's 35 years old?
I don't know how/why you came up with such strange idea.
x86 and ARM are 30-40 years old, with their widespread market share, such long history became an advantage as numerous tools & apps got built in that 30-40 years timeframe. it is called the ecosystem.
POWER as a failed ISA has been on its dead man walking stage for 30-40 years, nothing really get promoted & adopted. when you look back, it provides very strong certainty that if something good didn't happen in the last 30 years then there won't be a good chance for it happen in the next 5-10 years. its 30-40 years history is a disadvantage, because it is potential has been proven to be pretty bad.
This is like being 50 years old & highly successful in tech putting you in a comfort spot as your age implies your decades long valuable & successful experience. Being 50 years old with a shitty CV full of failures is a different story, your age is a huge liability for obvious reasons.
If the above is not obvious and you are in tech, then I have to say that you are in the wrong business.
SPARC, Intel i960, AMD29k, ARM, RISC-V and many others are direct descendants of the Berkeley RISC design.
Using the number of years of a measure of technological relevance, it can be argued that ARM and RISC-V designs are even older than 35 years. It is pure sophism, though, and it does not help.
By the same token, all modern computing architectures are fossilised designs because they are based on the von Neumann architecture from 1945.