numr 0.5.1

High-performance numerical computing with multi-backend GPU acceleration (CPU/CUDA/WebGPU)
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
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
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105

//! CUDA runtime tensor operation implementations
//!
//! This module loads all operation implementations from the src/ops/cuda/ directory.
//! Each operation type (type conversion, complex, normalization, etc.) is implemented in its own module.

use super::super::{CudaClient, CudaRuntime};
use crate::ops::TensorOps;

// TensorOps is a supertrait that composes all individual operation traits.
// Since we implement all the component traits in separate files, we just need
// this empty impl to satisfy the supertrait requirement.
impl TensorOps<CudaRuntime> for CudaClient {}

// Load all CUDA operation implementations from src/ops/cuda/
#[path = "../../../ops/cuda/type_conversion.rs"]
mod type_conversion;

#[path = "../../../ops/cuda/complex.rs"]
mod complex;

#[path = "../../../ops/cuda/normalization.rs"]
mod normalization;

#[path = "../../../ops/cuda/matmul.rs"]
mod matmul;

#[path = "../../../ops/cuda/conv.rs"]
mod conv;

#[path = "../../../ops/cuda/cumulative.rs"]
mod cumulative;

#[path = "../../../ops/cuda/activation.rs"]
mod activation;

#[path = "../../../ops/cuda/binary.rs"]
mod binary;

#[path = "../../../ops/cuda/unary.rs"]
mod unary;

#[path = "../../../ops/cuda/indexing/mod.rs"]
mod indexing;

#[path = "../../../ops/cuda/statistics.rs"]
mod statistics;

#[path = "../../../ops/cuda/random.rs"]
mod random;

#[path = "../../../ops/cuda/advanced_random.rs"]
mod advanced_random;

#[path = "../../../ops/cuda/quasirandom.rs"]
mod quasirandom;

#[path = "../../../ops/cuda/linalg.rs"]
mod linalg;

#[path = "../../../ops/cuda/reduce.rs"]
mod reduce;

#[path = "../../../ops/cuda/shape.rs"]
mod shape;

#[path = "../../../ops/cuda/sorting.rs"]
mod sorting;

#[path = "../../../ops/cuda/conditional.rs"]
mod conditional;

#[path = "../../../ops/cuda/utility.rs"]
mod utility;

#[path = "../../../ops/cuda/distance.rs"]
mod distance;

#[path = "../../../ops/cuda/multivariate.rs"]
mod multivariate;

#[path = "../../../ops/cuda/gemm_epilogue.rs"]
mod gemm_epilogue;

#[path = "../../../ops/cuda/semiring_matmul.rs"]
mod semiring_matmul;

#[path = "../../../ops/cuda/compare.rs"]
mod compare;

#[path = "../../../ops/cuda/scalar.rs"]
mod scalar;

#[path = "../../../ops/cuda/logical.rs"]
mod logical;

#[path = "../../../ops/cuda/einsum.rs"]
mod einsum;

#[cfg(feature = "fp8")]
#[path = "../../../ops/cuda/fp8_matmul.rs"]
mod fp8_matmul;

#[cfg(feature = "sparse")]
#[path = "../../../ops/cuda/sparse_24.rs"]
mod sparse_24;