61–70 of 312 posts
It seems like every 2 months I feel the burn of JS not having more standard primitive types and choices for numbers. I get this urge to learn Rust or Swift or Go which lasts about 15 minutes... until I realize how tied up I am with JS. But I do think one day (might take a while) JS will no longer be the obvious choice for front-end browser development.
In addition to the existing doubles, ES2020 added support for signed integers.
If anybody else was curious, it appears that the performance win of use of this instruction looks to be about 1-2% in general javascript workloads: https://bugs.webkit.org/show_bug.cgi?id=184023#c24
Is that significant enough to justify the instruction?
Unless I misread the current arm docs, I don't think this is still present in the ISA as of 2020? The whole RISC/CISC thing is long dead anyway, so I don't really mind having something like this on my CPU. Bring on the mill (I don't think it'll set the world on fire if they ever make it to real silicon but it's truly different)
What does it mean for the RISC/CISC thing to be dead? The distinction between them is more blurred than it used to be?
IBM's CPU-advancements of pipelining, out-of-order execution, etc. etc. were all implemented into Intel's chips throughout the 90s. Whatever a RISC-machine did, Intel proved that the "CISC" architecture could follow suite.
------
From a technical perspective: all modern chips follow the same strategy. They are superscalar, deeply-pipelined, deeply branch predicted, micro-op / macro-op fused "emulated" machines using Tomasulo's algorithm across a far larger "reorder buffer register" set which is completely independent of the architectural specification. (aka: out-of-order execution).
Ex: Intel Skylake has 180 64-bit reorder buffer registers (despite having 16 architectural registers). ARM A72 has 128-ROB registers (depsite having 32-architectural registers). The "true number" of registers of any CPU is independent of the instruction set.
I think it's best understood under a generalized ethnographic "utterance" aka linguistic and cultural problem in say a first order generalized theory of computing or combinatorics. aka the discrete log encoding::decode problem or the more ancient wisdom not passing down quick enough problem. When 2 group of static people of different cultures that some static group can just (lua) = implicitly define stuff very hard (combinatorical optimization) = branch and bound, it's better to remember than coding=calculate can be just thought as just a data structure problem (do a one time pad) in a generalized calculation in a automorphism and using techniques found in that abstract syntax tree that JavaScript can gradually typed more safely using TypeScript which Microsoft "already" gradually feed a generalized search engine (Google) knows so they can just Wikipedia (autocode the "Wisdom of the Crowds") it.
I'd understand it by understanding first principles from logic and congitive sciences and applied software engineering.
Software defined dynamic by necessity requires regularization to be very secure cryptographically to be ever exist in the first place.
But remember you can think about it anything another way. This is basically doing a reinterpret_cast in classical first order C++.
I personally think about it best as myself as a lexically scoped optimizing compiler using words compiled in some interior language to do a 1:1 mapping in Visual Studio to static analysis where my "utterances" can really apparently convince people that pattern matching stuff can really understood various matching techniques to recognize that these are just the same patterns.
Basic elementary physics and elementary internet-networking things to do figure out some paradoxes:
1. why aren't people focusing on general energy scavenging in principle? Apparently other people are. But it requires change of focus in many respects. Basically you can on a computer on software do it safely, but in the physical world, you may accidentally be doing very non-sensical things like imagining that you can just do an inversion of gravity.
2. remember that bias exists, but history suggests that bias can be thought many different ways. your experiences bias you, but sometimes humans beings need to think automatically in new adductive reasoning inferential systems in safety. That can only be done to affect the real world in computer simulation. This can add to human knowledge though. Deductive and inductive and coinductive can be performed over arbitrary domains in whatever order theoretic ways. 3. remember these are just analogies. They can condition your mind, but could also let your brain think more clearly. 4. for example implicit::ego::bias (abc) problem. 5. everyone who knows signal processing knows singular "poles" as 0 can be patched up by more modern stable manifold theory but only people who understand that boundaries are just arbitrary of linear spaces.
Remember than in the cultural sense of philosophy you can uw "hard" stuff once you understand traditionally thought of random [0,1] as boolean as more rich simplexes in the newer norm space [-1,1] which modern probability theory can be applied to. It creates a generalized system of norms of measurement of classical informational theory space.
Sorry for being so cryptic but recall that you could understand that shannon for example "accidentally" created the information by creating information and commutation theory by autodualizing math and a implicit association of a simple easy to read 1 page document that created a "digital bit". Equated as entropy. You could understand say hibert who let his "school" understand inf dimensional spaces.
I'll just point out a few free hugs in unicode spectrum (see https://home.unicode.org/emoji/emoji-frequency/). I just understand to look at it using free glue guns but people will have to discover the right modulation schemes themselves in discrete logarithms but it may require say, for example, doing 3 simultaneous generalizations to combine 2 y-combinators of information theoretic boolean bits could be remodulated into L-R folded "cultural" to use experience to debias past implicit bias and use + forward thinking bias to auto create quasi-well ordered programs that computer science people have already created.
let's just say I know by experience I know these things:
1. the www was originally created for people to collaborate over long distances to communicate more easily with clarity by being disconnected 2. social media was created for people to communicate as "new" ways but things facebook and twitter got too polluted by disinformation 3. telecomm systems eventually went from very messy error connecting codes modems and progressed to broadband by easing collaboration yet censorship still exists 4. search engine systems consuming more and more information to help locate stuff, but things like google get new bits of interformation to associations in their own internal db 5. it was at one time very hard to write international software but the invention of unicode helped people get access to the internet for the first time 6. the open/free source software like github and cultural practices like code reviews, allowed software developers for people around teams and corporations in remote areas to collaborate easily 7. computing devices like phones and smarter free computer languages allowed safer gradually and shifting codes of conduct to allow more open collaboration 8. the creation of cryptographic two session protocols allowed two parties security agree to securely exchange two exchange metadata together and work together 9. in applied software engineering, the creation of distributed version control systems like git and linux allowed mass adoption of free software 10. so these days in the software world you can just code software just by "coding" by pasting in semantic meaningful this is compositional software. 11. the productivity improvements in computer science has been enormous so that most people can save on energy both in hardware and software and refocus on more important stuff
Earlier quoted context omitted.
To understand RISC, ignore the acronym it stands for, instead just think fixed-instruction, load-store architecture. That's what RISC really means today. No variable-length instructions. No arithmetic instructions that can take memory operands, shift them, and update their address at the same time.
Someone once explained it like “not a reduced set, but a set of reduced instructions”. Not r(is)c, but (ri)sc. Pretty much what you say, I just liked the way of describing it.
If anybody else was curious, it appears that the performance win of use of this instruction looks to be about 1-2% in general javascript workloads: https://bugs.webkit.org/show_bug.cgi?id=184023#c24
Is that significant enough to justify the instruction?
Unless I misread the current arm docs, I don't think this is still present in the ISA as of 2020? The whole RISC/CISC thing is long dead anyway, so I don't really mind having something like this on my CPU. Bring on the mill (I don't think it'll set the world on fire if they ever make it to real silicon but it's truly different)
What does it mean for the RISC/CISC thing to be dead? The distinction between them is more blurred than it used to be?
It seems like every 2 months I feel the burn of JS not having more standard primitive types and choices for numbers. I get this urge to learn Rust or Swift or Go which lasts about 15 minutes... until I realize how tied up I am with JS. But I do think one day (might take a while) JS will no longer be the obvious choice for front-end browser development.
>But I do think one day (might take a while) JS will no longer be the obvious choice for front-end browser development. I think that day might be sooner than anyone thinks- Chromium is dominant enough now that their including Dart as a first-class language (or more likely, a successor to Dart) will likely be a viable strategy soon. Of course, the wildcard is Apple, but ultimately Dart can compile down to JS- being ab…
My guess is that this is what JIT people asked for.
Earlier quoted context omitted.
What does it mean for the RISC/CISC thing to be dead? The distinction between them is more blurred than it used to be?
RISC / CISC was basically IBM-marketing speak for "our processors are better", and never was defined in a precise manner. The marketing is dead, but the legend lives on years later. IBM's CPU-advancements of pipelining, out-of-order execution, etc. etc. were all implemented into Intel's chips throughout the 90s. Whatever a RISC-machine did, Intel proved that the "CISC" architecture could follow suite. ------ From a t…
Why wasn't it obvious previously? A few things had to happen: compilers had to evolve to be sophisticated enough, mindsets had to adapt to trusting these tools to do a good enough job (I actually know several who in the 80' still insisted on assembler on the 390), and finally, VLSI had to evolve to the point where you could fit an entire RISC on a die. The last bit was a quantum leap as you couldn't do this with a "CISC" and the penalty for going off-chip was significant (and has only grown).