ternlang-core 0.3.3

Compiler and VM for Ternlang — balanced ternary language with affirm/tend/reject trit semantics, @sparseskip codegen, and BET bytecode execution.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

// std::tensor — ternary tensor operations
// RFI-IRFOS Ternary Intelligence Stack
// use std::tensor;

// Allocate a 1D ternary tensor of given size (all hold/zero initialised)
fn zeros(size: int) -> trittensor<1 x 1> {
    let t: trittensor<1 x 1>;
    return t;
}

// Sparse matrix multiply — delegates to @sparseskip built-in
// This is the flagship operation of the ternary stack
fn sparse_mm(a: trittensor<1 x 1>, b: trittensor<1 x 1>) -> trittensor<1 x 1> {
    @sparseskip let result: trittensor<1 x 1> = matmul(a, b);
    return result;
}

// Dense matrix multiply — no zero skipping
fn dense_mm(a: trittensor<1 x 1>, b: trittensor<1 x 1>) -> trittensor<1 x 1> {
    let result: trittensor<1 x 1> = matmul(a, b);
    return result;
}