matrixmultiply 0.3.10

General matrix multiplication for f32 and f64 matrices. Operates on matrices with general layout (they can use arbitrary row and column stride). Detects and uses AVX or SSE2 on x86 platforms transparently for higher performance. Uses a microkernel strategy, so that the implementation is easy to parallelize and optimize. Supports multithreading.
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
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

// Copyright 2020, 2022 Ulrik Sverdrup "bluss"
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.

use rawpointer::PointerExt;

/// A Send + Sync raw pointer wrapper
#[derive(Copy, Clone)]
#[repr(transparent)]
pub(crate) struct Ptr<T> { ptr: T }
unsafe impl<T> Sync for Ptr<*const T> { }
unsafe impl<T> Sync for Ptr<*mut T> { }
unsafe impl<T> Send for Ptr<*const T> { }
unsafe impl<T> Send for Ptr<*mut T> { }

/// Create a Ptr
///
/// # Safety
///
/// Unsafe since it is thread safety critical to use the raw pointer correctly.
#[allow(non_snake_case)]
pub(crate) unsafe fn Ptr<T>(ptr: T) -> Ptr<T> { Ptr { ptr } }

impl<T> Ptr<T> {
    /// Get the pointer
    pub(crate) fn ptr(self) -> T
        where T: Copy
    {
        self.ptr
    }
}

impl<T> Ptr<*mut T> {
    /// Get as *const T
    pub(crate) fn to_const(self) -> Ptr<*const T> {
        Ptr { ptr: self.ptr }
    }
}

impl<T> PointerExt for Ptr<*const T> {
    #[inline(always)]
    unsafe fn offset(self, i: isize) -> Self {
        Ptr(self.ptr.offset(i))
    }
}

impl<T> PointerExt for Ptr<*mut T> {
    #[inline(always)]
    unsafe fn offset(self, i: isize) -> Self {
        Ptr(self.ptr.offset(i))
    }
}