Thrust Programming Language

General Characteristics

Why Thrust?

asmfn invoke_x86_64_exit_syscall() void@asmSyntax("AT&T") @convention("C"){"mov $$60, %rax","mov $$1, %rdi","syscall"}{"~{rax}~{rdi}"}fn main() s32 @public{    invoke_x86_64_exit_syscall();return0;}

Current Features

Future Features

State

Documentation

Examples

Compiler

Linux

./thrustc -opt=O3 fibonacci.thrust -start -o fibonacci -end && ./fibonacci

Windows

.\thrustc.exe -opt=O3 fibonacci.thrust -start -o fibonacci.exe -end && .\fibonacci.exe

Package Manager

torio run

Code Example - Hello World

// ******************************************************************************************////   Hello World!//// ******************************************************************************************// Thrust Programming Language - File extensions//// - '.🐦'// - '.Thrust'//// External declaration for the C printf functionfn print(fmt:const array[char]) s32 @public @arbitraryArgs @extern("printf") @convention("C");fn main() s32 @public{    print("Hello World!");return0;}

Code Example - Fibonacci

// ******************************************************************************************////   Fibonacci - O(2^n)//// ******************************************************************************************// Thrust Programming Language - File extensions// // - '.🐦'// - '.thrust'//// External declaration for the C printf functionfn print(fmt:const array[char]) s32 @public @arbitraryArgs @extern("printf") @convention("C");// Computes the nth Fibonacci number recursively//// Parameters://   n: The index of the Fibonacci number to compute (unsigned 32-bit integer)//// Returns: The nth Fibonacci number (unsigned 32-bit integer)//// Attributes://   @hot: Marks the function as frequently executed, encouraging aggressive optimizations.//   @inline://         Maps to LLVM's 'inlinehint', suggesting the compiler inline this function//         at call sites to reduce call overhead. May increase code size, and may be//         limited for deep recursion.//   @nounwind://         Maps to LLVM's 'nounwind', guaranteeing the function will not unwind the stack//         (i.e., it will not throw an exception or cause an abnormal termination//         that requires stack unwinding). This enables significant optimizations.//fn fibonacci(n:u32) u32@hot@inline@nounwind{   if n <=1{       return n;   }   return fibonacci(n -1) + fibonacci(n -2);}// Prints the first n Fibonacci numbers// Parameters://   n: The number of Fibonacci numbers to print (unsigned 32-bit integer)fnprintFibonacci(n:u32) void{   for localmuti:u32 =0; i <n; ++i;{       print("%d\n",fibonacci(i));   }}fnmain(argc:u32,argv: ptr[array[char]]) s32 @public{    print("Fibonacci sequence: ");             printFibonacci(25);               return0;}

Code Example - 100 Millions Array Updates

// ******************************************************************************************////   100 MILLIONS ARRAY UPDATES//// ******************************************************************************************fn atoi(str:const array[char]) s32 @public @vaArgs @extern("atoi") @convention("C");fn srand(seed:u32) void @public @extern("srand") @convention("C");fntime(timer:ptr) u32 @public @extern("time") @convention("C");fn rand() s32 @public @extern("rand") @convention("C");fnprint(fmt:const array[char]) s32 @public @arbitraryArgs @extern("printf") @convention("C");fn main(argc:s32,argv: ptr[array[char]]) s32 @public{    local mut u: s32 =atoi((deref(argv[1]asptr[ptr[char]]))asconstarray[char]);srand(time(nullptr));       local mut r: s32 =rand() %10000;     local mut a: array[s32;10000];for localmut i: s32 =0; i <10000;i++;{for localmut j: s32 =0; j <100000;j++;{            a[i] =(deref a[i]) +((j % u));}        a[i] =(deref a[i]) + r;}print("%ld\n", deref a[r]);return0;}

#include"stdio.h"#include"stdlib.h"#include"stdint.h"#include"time.h"intmain(intargc,char**argv) {intu=atoi(argv[1]);srand(time(NULL));intr=rand() %10000;int32_ta[10000]= {0};for (inti=0;i<10000;i++) {for (intj=0;j<100000;j++) {a[i]=a[i]+j%u;                      }a[i]+=r;                            }printf("%d\n",a[r]);              }

Benchmark

Note

Actually, it can be a margin of error and is the same as C, although with-jit it even outperforms C, even though the Just-In-Time Compiler always has a heavy overhead at startup.

Thrust

thrustc -opt=O3 loop.thrust -start -o loop -end && ./loop 1

C

clang -O3 loop.c -o loop && ./loop 1

Note

Obviously, if you have a little knowledge of CS, you know that this isn't the ideal way to test which programming language is faster, but anyway, it's just to point out that Thrust is trying to be a C equivalent in the speed field.

Case Of Study

Background

Support

Social Networks

Always Remember