redstone-ml 0.0.0

High-performance Machine Learning, Auto-Differentiation and Tensor Algebra crate for Rust
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

use crate::flat_index_generator::FlatIndexGenerator;
use crate::impl_default_trait_for_dtypes;
use crate::ndarray::collapse_contiguous::collapse_to_uniform_stride;
use crate::{define_binary_op_trait, simd_binary_op_specializations};
use paste::paste;
use std::ops::Sub;
use std::ptr::addr_of;

define_binary_op_trait!(BinaryOpSub, Sub, sub, -;
                        i8, i16, i32, i64, i128, isize,
                        u8, u16, u32, u64, u128, usize);

impl BinaryOpSub for f32 {
    simd_binary_op_specializations!(sub);

    #[cfg(all(apple_vdsp, not(neon_simd)))]
    unsafe fn sub_stride_n_n(lhs: *const Self, lhs_stride: usize,
                             rhs: *const Self, rhs_stride: usize,
                             dst: *mut Self, count: usize) {
        use crate::acceleration::vdsp::vDSP_vsub;
        vDSP_vsub(rhs, rhs_stride as isize, lhs, lhs_stride as isize, dst, 1, count);
    }
}

impl BinaryOpSub for f64 {
    simd_binary_op_specializations!(sub);

    #[cfg(all(apple_vdsp, not(neon_simd)))]
    unsafe fn sub_stride_n_n(lhs: *const Self, lhs_stride: usize,
                             rhs: *const Self, rhs_stride: usize,
                             dst: *mut Self, count: usize) {
        use crate::acceleration::vdsp::vDSP_vsubD;
        vDSP_vsubD(rhs, rhs_stride as isize, lhs, lhs_stride as isize, dst, 1, count);
    }
}