ferrum-kernels 0.7.7

Unified compute kernels (CUDA/Metal/CPU) and model runner for Ferrum inference
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
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176

/***************************************************************************************************
 * Copyright (c) 2023 - 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 * SPDX-License-Identifier: BSD-3-Clause
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *
 * 1. Redistributions of source code must retain the above copyright notice, this
 * list of conditions and the following disclaimer.
 *
 * 2. Redistributions in binary form must reproduce the above copyright notice,
 * this list of conditions and the following disclaimer in the documentation
 * and/or other materials provided with the distribution.
 *
 * 3. Neither the name of the copyright holder nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 **************************************************************************************************/
#pragma once

#include <cute/config.hpp>                   // CUTE_HOST_DEVICE
#include <cute/pointer_base.hpp>             // cute::iter_adaptor
#include <cute/swizzle.hpp>                  // cute::Swizzle, cute::get_swizzle primary template
#include <cute/util/type_traits.hpp>         // cute::iterator_traits
#include <cute/container/array_subbyte.hpp>  // cute::subbyte_iterator

/* This implements a swizzle pointer of the form
 *   InvolutionFn o PtrAdd
 * where the InvolutionFn need not be linear.
 *
 * This differs subtly from swizzle_layout because the smem pointer is used
 * as the offset. That means that swizzle_layout will implement position-independent
 * swizzle layouts, while swizzle_ptr implements position-dependent swizzle tensors.
 * Arch chose to design hardware with position-dependent swizzles.
 *
 * For clarity:
 *   NormalLayout  : DeRef <- PtrAdd <- [Layout]
 *   ComposedLayout: DeRef <- PtrAdd <- [Swizzle <- OffsetAdd <- Layout]
 *   SwizzlePtr    : [DeRef <- Swizzle <- PtrAdd] <- Layout
 *
 * Furthermore, for known swizzles, this pointer attempts to decay itself
 *    to a normal-pointer with a new layout containing dynamic or static strides.
 * This is possible by determining the subdomain of the InvolutionFn
 *    that is identity and testing if the Layout's codomain is contained
 *    within it.
 */

namespace cute
{

// concept SwizzleFn {
//   CUTE_HOST_DEVICE constexpr static uint apply(uint);
// }
// See Swizzle<B,M,S> in swizzle.hpp for common swizzle-functions.

template <class SwizzleFn, class Iterator>
struct swizzle_ptr : iter_adaptor<Iterator,swizzle_ptr<SwizzleFn,Iterator>>
{
  using iterator     = Iterator;
  using reference    = typename iterator_traits<iterator>::reference;
  using element_type = typename iterator_traits<iterator>::element_type;
  using value_type   = typename iterator_traits<iterator>::value_type;

  using iter_adaptor<Iterator,swizzle_ptr<SwizzleFn,Iterator>>::iter_adaptor;

  template <class Iter>
  CUTE_HOST_DEVICE constexpr static
  Iter apply_swizzle(Iter ptr) {
    return {apply_swizzle(ptr.get())};
  }

  template <class T>
  CUTE_HOST_DEVICE constexpr static
  T* apply_swizzle(T* ptr) {
    return reinterpret_cast<T*>(SwizzleFn::apply(reinterpret_cast<uintptr_t>(ptr)));
  }

  template <class T>
  CUTE_HOST_DEVICE constexpr static
  subbyte_iterator<T> apply_swizzle(subbyte_iterator<T> ptr) {
    return {apply_swizzle(ptr.ptr_), ptr.idx_};
  }

  CUTE_HOST_DEVICE constexpr
  reference operator*() const {
    return *apply_swizzle(this->get());
  }

  template <class Int>
  CUTE_HOST_DEVICE constexpr
  reference operator[](Int const& i) const {
    return *apply_swizzle(this->get() + i);
  }
};

//
// Helper Function
//
template <class SwizzleFn, class P>                   // Found the SwizzleFn
struct get_swizzle<swizzle_ptr<SwizzleFn,P>> { using type = SwizzleFn; };
template <class T>                                    // Recurse into anything with a ::iterator
struct get_swizzle<T, void_t<typename T::iterator>> : get_swizzle<typename T::iterator> {};

template <class Iterator, class SwizzleFn>
CUTE_HOST_DEVICE constexpr
swizzle_ptr<SwizzleFn,Iterator>
make_swizzle_ptr(Iterator ptr, SwizzleFn) {
  return {ptr};
}

// Swizzle-0 specialization for immediate decay
template <class Iterator, int M, int S>
CUTE_HOST_DEVICE constexpr
Iterator
make_swizzle_ptr(Iterator ptr, Swizzle<0,M,S>) {
  return ptr;
}

//
// Recast
//

template <class SwizzleFn, class P>
CUTE_HOST_DEVICE constexpr
auto
raw_pointer_cast(swizzle_ptr<SwizzleFn,P> const& ptr) {
  return raw_pointer_cast(ptr.get());
}

// SwizzleFn operates on the pointer address, so it doesn't care about the type
template <class NewT, class SwizzleFn, class P>
CUTE_HOST_DEVICE constexpr
auto
recast_ptr(swizzle_ptr<SwizzleFn,P> const& ptr) {
  return make_swizzle_ptr(recast_ptr<NewT>(ptr.get()), SwizzleFn{});
}

// The statically-known alignment of a swizzle pointer is the alignment of the swizzle function converted to bits
template <class SwizzleFn, class P>
CUTE_HOST_DEVICE constexpr
auto
max_alignment(swizzle_ptr<SwizzleFn,P> const&) {
  return Int<8>{} * max_alignment(SwizzleFn{});
}

//
// Display utilities
//

template <class SwizzleFn, class P>
CUTE_HOST_DEVICE void print(swizzle_ptr<SwizzleFn,P> ptr)
{
  print(SwizzleFn{}); printf("_"); print(ptr.get());
}

#if !defined(__CUDACC_RTC__)
template <class SwizzleFn, class P>
CUTE_HOST std::ostream& operator<<(std::ostream& os, swizzle_ptr<SwizzleFn,P> ptr)
{
  return os << SwizzleFn{} << "_" << ptr.get();
}
#endif

} // end namespace cute