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
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
/*
* Copyright (c) Microsoft Corporation.
* Licensed under the MIT license.
*/
//! # GEMM-lite Matrix Kernels
//!
//! This module is a work in progress. There are many pieces that are still needed:
//!
//! * Proper run time cache-size detection.
//! * GEMM-like kernels beyond "maxsim".
//! * Quantization support.
//! * Blocking along the contraction dimension "k".
//! * Comprehensive performance tuning.
//!
//! For consistency with GEMM terminology, the following conventions are used.
//!
//! A matrix-kernel has the general form
//!
//! ```text
//! C[M x N] = A[M x K] . B[K x N]
//! ```
//! with internal dimensions as follows:
//!
//! * `M`: The number of rows of `A` and `C`.
//! * `N`: The number of columns of `B` and `C`.
//! * `K`: Contraction dimension. Columns of `A` and rows of `B`.
//! * `MR`: Packing parameter for `A`. This is the number of rows processed in a micro-kernel.
//! * `NR`: Packing parameter for `B`. This is the number of columns processed in a micro-kernel.
use NonZeroUsize;
// Kernels
pub
// Arguments
pub
// private
// re-export
pub use Drive;
pub use DimK;
/// Placeholder model for CPU caches.
pub
//------------//
// Test utils //
//------------//