Function einsum_with_plan_into 

pub fn einsum_with_plan_into<T: Scalar + HasAlgebra>(
    tree: &ContractionTree,
    operands: &[&Tensor<T>],
    alpha: T,
    beta: T,
    output: &mut Tensor<T>,
) -> Result<()>

Execute einsum with a pre-optimized ContractionTree, accumulating into an existing output.

Computes output = alpha * einsum(operands) + beta * output. Avoids both subscript parsing and contraction order optimization. This is the fastest variant for hot loops with pre-allocated buffers.

Examples

use tenferro_einsum::{einsum_with_plan_into, ContractionTree, Subscripts};
use tenferro_tensor::{Tensor, MemoryOrder};
use tenferro_device::LogicalMemorySpace;

let col = MemoryOrder::ColumnMajor;
let subs = Subscripts::new(&[&[0, 1], &[1, 2]], &[0, 2]);
let tree = ContractionTree::optimize(&subs, &[&[3, 4], &[4, 5]]).unwrap();
let mut c = Tensor::<f64>::zeros(&[3, 5], LogicalMemorySpace::MainMemory, col);

// Hot loop: reuse output buffer, no allocation per iteration
for _ in 0..1000 {
    einsum_with_plan_into(&tree, &[&a, &b], 1.0, 0.0, &mut c).unwrap();
}

Errors

Returns an error if the operand shapes do not match those used to build the contraction tree, or the output shape is incorrect.