The initial conception of a product or service design is often ideal and beautiful. However, for that product or service to be used in the real world, changes and additions must be made to accommodate different requirements and actual usage conditions. The same is true for computer design. Even if a designer comes up with an ideal design, its beauty is lost in the process of improvement based on its actual operation and feedback from users.
This is not necessarily a bad thing. Rather, it is a sign that the product or service is being used in the real world and making a contribution to people’s lives. The pursuit of practicality can also lead to new innovations.
Thus, a gap exists between idealistic design and real-world operation, and successfully bridging this gap will lead to the success of the product or service. It is important to have a pragmatic viewpoint and to pursue practicality as well as beauty.
@tanakh: I mean, I think RISC-V had some sort of Packed SIMD Extension, but I wonder how it’s segregated? In the RISC-V book, Isoter criticized Isoter for adding a large number of instructions every time they changed the SIMD width, and said that with variable-length vector instructions, they didn’t need to do that.
@a4lg: @tanakh The Packed SIMD being developed for RISC-V uses general registers and aims at extensions suitable for, for example, DSPs. It is integer-only, GPR-only, etc. I think it is integer-only and GPR-only, so it can be separated from vector instructions to some extent.
ShougoMatsu This is not limited to CPUs. It is very possible that a beautiful design or architecture remains beautiful simply because no one uses it. When practicality is considered, more and more functions are added and it becomes cluttered and dirty. I could acutely realize that there is “no archie” in the world that is clean and convenient for all to use. … ShougoMatsu If RISC-V is getting dirtier, that’s desirable because that’s how much it’s used in the real world. ShougoMatsu The more people use my plug-ins, the less beautiful they are in the beginning. That’s just the way it is. ShougoMatsu So I would look at someone who emphasizes beauty with a cold eye. I thought I was living only in an ideal world. I am a realist.
@a4lg: I really agree with this. It would have been impossible for anyone to anticipate and create everything ahead of time, and RISC-V was not an example of that as a result. Well, I personally feel it’s still better, but still not the best and cleanest of everything.
prompt
Abstract from technical terms and make generalizations.
These tweets cover the topic of the RISC-V computer architecture, specifically its SIMD (Single Instruction, Multiple Data) extension. However, abstracting the technical terminology to a general discussion, we have the following…
Comments
.
- Well, people who say “the design is dirty, so I’ll clean it up” appear at a high frequency, but that doesn’t stop them from trying to make something new and becoming more knowledgeable in their field, which is beneficial.
- It is beneficial to experience the process of making something beautiful and then having it become dirty in the process of being used, so that people who can balance realism and aesthetics can be nurtured.
- It’s not useful to criticize people who are just criticizing dirty things and don’t try to make them themselves and think they’re winning by criticizing people who are trying to make clean things, saying that they don’t see reality, etc.
relevance - Get rid of critics who don’t take positions.
This page is auto-translated from /nishio/使われていないから綺麗 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.