>compiled languages are faster than interprted
why dont they make a version of x86 assembler that compiles?
>compiled languages are faster than interprted
why dont they make a version of x86 assembler that compiles?
You have no idea what you're talking about.
not an argument
Ugh, fine. There already is a technique that attempts to take interpreted code and make it native (not exactly compiling, but it's the closest thing to what you're describing).
It's called JIT (Just-In-Time) execution. Go take a look into it.
lol you're literaly retarded
Because?
AHAHAAHA
sorry dude you arent even at the intellectual level to discuss technology with a retarded chuuni 14 year old devil girl
gtfo out of this thread
But CPU instructions must be "interpreted", there's no other way to run a CPU
Unless you are talking about a "compiler" that turns assembly into a full-featured VHDL design and drops it on an FPGA or something
>But CPU instructions must be "interpreted", there's no other way to run a CPU
you have 15 seconds to prove this
This thread is fucking cancer
not an argument
This
Stop avatarfagging please
>JIT x86 asm
ahahaha retard
can't we just compile the machine code while we're at it?
its not avatar use
Brainlets getting btfo by a satania-san poster.
it is avatar use
Go fuck a hyena
no its not
not an argument retard
Your machine's CPU doesn't re-build itself for every single program that you run, it looks at each instruction, figures out what to do, and performs the necessary action. It's "interpreting" opcodes in the same way that the JVM "interprets" bytecode.
In a sense, "compiling" takes the final output one abstraction down. While JVM interprets bytecode, a compiler that generates native code effectively creates programs that reside on the same "abstraction level" as the JVM.
Therefore, for a truly "compiled" CPU, your compiler must go one layer lower than assembly, which is the CPU microarchitecture itself.
try again brainlet
dont bother replying
filtered
In interpreted languages you can do stuff doesn't map well to low level assembly
For example python classes. They don't have a static definition of what methods and members it has. You can add and remove members from a python object however you please. You even have metaclasses that can modify the instance before it's created. Because of those design choices everything that's inside a python object must be stored inside a dictionary to allow dynamic changes at runtime. Compiled or not, you still have to index a dictionary to interact with python objects. Also python is dynamically typed so everything must be a reference, even a basic integer.
In C++ you cannot change the structure of an object at runtime so it can have a fixed memory layout which allows fast addressing. C++ is also statically typed and so it can work with unboxed values.
Reminder to sage and report
>sage doesn't show anymore
Why bother? Just post "shit thread" and it will do.
hmm hmm very interesting very interesting
HOWEVER....
>"b-but muh cpu actually translates cisc to a risc subset under the hood"
Yeah but it's still interpreting that subset, you fucking clown
Can I lick your ass clean?
gentoo does this by default
...
Anything below machine code is the metal. The signals running through are based on electrical gates. This is not interpretation.
is there a way to manually switch gates using an electrode? this is the only way to code for me. reminds me of the old days.