par_iter_mut_simd

fn par_iter_mut_simd(x: &mut Tensor<T>) -> ParStridedMutSimd<T>

similar as par_iter_simd, input pass to the closure will be mutable. You can use it to do inplace computation.

Parameters:

x: Tensor to iterate

Returns:

ParStridedMutSimd

Note

Simd closure will only be execute when the result simd type number of lane is equal to input's.
In Simd closure, result can only call write_unaligned method to write data.

Examples:

use hpt::iter::{ParStridedIteratorSimd, TensorIterator};
use hpt::types::math::FloatOutUnary;
use hpt::Tensor;

fn main() -> anyhow::Result<()> {
    let mut x = Tensor::<f64>::new(&[1f64, 2., 3.]);

    x.par_iter_mut_simd().for_each(
        |x| {
            *x = x.sin();
        },
        |x| {
            let val = x.read_unaligned();
            x.write_unaligned(val._sin());
        },
    );

    println!("{}", x);
    Ok(())
}

Backend Support

Backend	Supported
CPU	✅
Cuda	❌