 |
Post Au93JJOyNzt4fDYDmC by [email protected] |
 |
More posts by [email protected] |
|
Post #Au7UzQJbNiMQiKQKLg by [email protected] |
|
0 likes, 4 repeats |
|
|
|
You might not like it, but this is what peak compiler performance looks like. |
|
|
|
Post #Au7b4RHhnurMtSJN9k by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela how can the compiler break down a recursive function to a mathematical sta… |
|
|
|
Post #Au7b4ROROsFPELSkee by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@viduq It's not exactly brute force, but it does involve some heavy lifting… |
|
|
|
Post #Au8lQ6LbWPfT1q3yDY by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela This is brilliant! I'd like to put this into my uni lecture on C/C++. … |
|
|
|
Post #Au8lQ6Rz8glvLd34AC by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@josch This code snipped has been floating around for a while, here's a blo… |
|
|
|
Post #Au8tnTcJyHJJUYXVOS by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela Unfortunately this does not work for the mutually recursive isEven and isO… |
|
|
|
Post #Au8tnTihaYPloLWbL6 by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@realleif Inline harder! |
|
|
|
Post #Au8uOueiPBK646VgrQ by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela Actually, this is what bad code writing looks like. |
|
|
|
Post #Au8uOulnyoziQ5pLua by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@caravantraveller This might be an extreme case of bad code, yes. But the moral… |
|
|
|
Post #Au92jDlxq0UumR5yts by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela This is quite impressive, and I wonder how complex you could make these ki… |
|
|
|
Post #Au92jDtlN0jhAckD3Y by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@loke @ela I would expect that this is the combination of a few things:First, t… |
|
|
|
Post #Au939I9w9Ps9ukjwci by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @viduq Uh, I need Alt text to help me understand this.😢 |
|
|
|
Post #Au93GVpXijkkLEdz28 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela this stumped me until I saw clang decides to return bool in al, not eax, f… |
|
|
|
Post #Au93GW0B5CGAsDcTbs by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@mirabilos @ela does "and al,1" clear the upper bits? |
|
|
|
Post #Au93JJOyNzt4fDYDmC by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela wild. |
|
|
|
Post #Au93JLljaLHa0ODQX2 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@waldi @ela only seven of them |
|
|
|
Post #Au93JLtB8fEmNThN8S by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@mirabilos @waldi To be unfathomably pedantic, that would be all of them for al. |
|
|
|
Post #Au93JM0yffTYlfLbI8 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @waldi yes, but it using only al ipv eax for the return value strikes me a… |
|
|
|
Post #Au93JM8QDzQl8kpXtY by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@mirabilos @waldi Well, it's what the spec says. |
|
|
|
Post #Au93MDdOHmNBe3gHEu by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela It seems rustc is doing the same thing: |
|
|
|
Post #Au93MDpnXeIWGXUBZw by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@sergi @ela LLVM ftw |
|
|
|
Post #Au93N28eVp5YHUgwuO by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela “The computers rebelled but quietly and very performantly so no one noti… |
|
|
|
Post #Au93NJi1BaTkm6MvtQ by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela I love how the ALT text really explains what's going on even for peopl… |
|
|
|
Post #Au93NJqAhH07BOBRbM by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@tante @ela thanks for the hint, wouldn't have looked at it otherwise. |
|
|
|
Post #Au93ZN8mFkjotPdLoO by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@viduqIt uses an LLM to explain the code (no I'm joking, we're not ther… |
|
|
|
Post #Au93bxFMyv53SBwsIi by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela no way... |
|
|
|
Post #Au93bxQMK3s40H5eQi by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@plusmid Way! Try for yourself. |
|
|
|
Post #Au93cUJm1ZDA1PGMFM by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela This is INSANE .. I mean it's insane the C code written could not be w… |
|
|
|
Post #Au93cURvXFjWQh4rxI by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@gilesgoat @ela, it's not “i % 2”, as that would give an inverted resul… |
|
|
|
Post #Au93cUa52wFspytNfE by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@lp0_on_fire @ela yes but if you want an isODD(x) that returns '1' if i… |
|
|
|
Post #Au93f6otqUMo0r4VFo by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela I wrote a blog post years ago that shows all the intermediate steps of per… |
|
|
|
Post #Au93fvkOV049xeM7oO by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela Nice! I wonder how this feat is performed? I suppose the tail-call is elim… |
|
|
|
Post #Au93fwLyFI7jqBw8Nk by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela ah whelp u answered this while I was typing :'Dhttps://infosec.exchang… |
|
|
|
Post #Au93oOHqH2Amn82EhE by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela Interestingly clang does not think that "bool isEven(unsigned int i) … |
|
|
|
Post #Au93vWWg2SfUN8NKBU by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ljrk @ela the truly amazing thing about compiler explorer with clang is, that … |
|
|
|
Post #Au93yjanujDBlYpvkG by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @naoEhAlexa |
|
|
|
Post #Au940itrSaXYCHhO1g by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela phew, at least the Go compiler is calmingly unoptimized 😅 |
|
|
|
Post #Au9418FLDrZoo8wOTg by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@porglezomp @ela very cool. I have no idea about compiler internals but were ab… |
|
|
|
Post #Au941WKpf6EHW7h7Mu by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@plusmid There's more than one go with llvm backend project. :) |
|
|
|
Post #Au943qBwo2gVgGe3Vo by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@viduq @ela Check out how examples like this are also reduced to simple polyno… |
|
|
|
Post #Au944H30xMa8yNfnpw by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@gilesgoat No, that's purely by analysis. |
|
|
|
Post #Au947DUGeOWUnbp8aG by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela holly shit... |
|
|
|
Post #Au948CctKwfOQ1KNzk by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@[email protected] so from now on write your isEven function recursive? Got … |
|
|
|
Post #Au94BCvQgK0sDNQoiG by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@lbehm @ela Wow, I didn't even know that! I knew that clang has this featur… |
|
|
|
Post #Au94BGaJ4nUfZ1a0LQ by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@mirabilos @ela Ah, because _Bool is defined as 8bit value where only bit 0 mat… |
|
|
|
Post #Au94BGgKiOJXriOojo by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@waldi @ela except for actual storage (where one bit would be better in structs… |
|
|
|
Post #Au94MlUcV5b4brP1yC by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela But then it is not fair to call the source code "very inefficient&quo… |
|
|
|
Post #Au94Qdmf6WdHrJgpPc by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@loke The problem here is that the compiler can't tell if something is just… |
|
|
|
Post #Au94RATfYA33Fy3ctE by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela A sane implementation of the function would result in more machine code, I… |
|
|
|
Post #Au94RTzuyU8LnUI9IG by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @viduq Black magic, but as a beauty. |
|
|
|
Post #Au94S0xaQAsQ1oSFk0 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela It's insane ( in a good way for the compiler ). |
|
|
|
Post #Au94TEogOAudICsDPk by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela there certainly is. I think TinyGo uses the LLVM under the hood. Would be … |
|
|
|
Post #Au94Vkp6MLwL8gGkj2 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@elaWow |
|
|
|
Post #Au94WeZR5mgQ2JN25A by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @viduq compiler design is so beautiful ♥️ |
|
|
|
Post #Au94YHxM8spUTkL3tA by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela Or maybe a raw bitshift to take the last digit and check if its 0,2,4,6 or… |
|
|
|
Post #Au94aqGHZDqjSCPPKC by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@Ehay2k @viduq alt text is right here: https://archive.org/details/advancedcomp… |
|
|
|
Post #Au94azzzMel3eYrXJg by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@loke @ela In general, this would be at least as hard as the halting problem. P… |
|
|
|
Post #Au94exGnS6mGLdbv1s by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela Love it! |
|
|
|
Post #Au94hWXdHhigJti6qW by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@waldi @mirabilos Indeed. |
|
|
|
Post #Au94siphvQXZcvVLN2 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela This is the “give them something better than they would choose to give t… |
|
|
|
Post #Au94wjQTf7JIcmcD7A by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela This is nice. I knew that C and C++ compiler pipelines and toolchains are … |
|
|
|
Post #Au9IpFoyjTjY8cP7Q0 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @waldi I admit only vague familiarity with the psABI for amd64 (normally p… |
|
|
|
Post #Au9JObPaS3s3EdNnAO by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @loke There's a lot of examples of this. Like mergesort vs quicksort (… |
|
|
|
Post #Au9Lr3wi9ga1kVZi40 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@ela @viduq 🤩 |
|
|
|
Post #AuB6KsqGnH7n32gCm0 by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@ela i have no idea what i'm saying but i did checked what this means.you p… |
|
|
|
Post #AuB6kVbfnybYVv4U88 by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@ela i have no idea what i'm saying but i did some research on it...We put … |
|
|
|
Post #AuB8D80OjUV1fTqYds by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@ela perhaps `AND` works this way when we compare digits:if 2 digits are equals… |
|
|
|
Post #AuBN1DOyLGcQF47Q3s by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@porglezomp @ela you start with the tail recursion optimization but here isEven… |
|
|
|
Post #AuBN1DXpoJhwgYGUsK by [email protected] |
|
0 likes, 1 repeats |
|
|
|
@nrab @porglezomp LLVM extends that basic algorithm a tad bit by trying to move… |
|
|
|
Post #AuUePfu5Pl8ioT9xmi by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@elaIt would be cool if the compiler could create a logfile where it explains a… |
|
|
|
Post #AuUePsx2u4EPHPe0hc by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@jaj It can! At least in the specific case of llvm/clang. |
