ruPRIM 0.1.1

Parallel primitives, reductions, scans, and indexing for Ruda.
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
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249

use ruda_kernel::dsl as kernel_dsl;
use ruda_kernel::dsl::Runtime;
use ruda_kernel::dsl::tensor_vector_size_parallel;
use ruda_kernel::tensor::layout::address_type;
use ruda_kernel::tensor::layout::shape_divmod;
use ruda_kernel::tensor::RudaTensor;
use ruda_kernel::dsl::calculate_ruda_count_elemwise;
use ruda_kernel::dsl::intrinsic;
use ruda_kernel::dsl::prelude::*;
use ruda_kernel::library::FastDivmod;
use ruda_kernel::library::tensor::layout::linear::LinearView;

#[ruda(launch_unchecked, address_type = "dynamic")]
fn slice_assign_kernel<E: Numeric, N: Size>(
    input: &mut Tensor<Vector<E, N>>,
    value: &LinearView<Vector<E, N>>,
    slice_shape: Sequence<FastDivmod<usize>>,
    slice_offsets: Sequence<usize>,
    #[define(E)] _dtype: StorageType,
) {
    if !value.is_in_bounds(ABSOLUTE_POS) {
        terminate!()
    }

    let rank = comptime!(slice_shape.len());

    let line_size = input.vector_size();
    let mut offset_remainder = ABSOLUTE_POS * line_size;
    let mut offset_input = 0;

    #[allow(clippy::explicit_counter_loop)]
    #[unroll]
    for i in 0..rank {
        let dim = rank - i - 1;
        let (rem, offset_local) = slice_shape[dim].div_mod(offset_remainder);

        let range_start = slice_offsets[dim];
        let offset_local_input = offset_local + range_start;

        offset_input += offset_local_input * input.stride(dim);
        offset_remainder = rem;
    }

    // Value tensor is accessed linearly since it's a LinearView
    input[offset_input / line_size] = value[ABSOLUTE_POS];
}

/// Kernel for slice assign with steps
#[ruda(launch_unchecked, address_type = "dynamic")]
fn slice_assign_with_steps_kernel<E: Numeric>(
    input: &mut Tensor<E>,
    value: &LinearView<E>,
    value_shape: Sequence<FastDivmod<usize>>,
    starts: Sequence<usize>,
    ends: Sequence<usize>,
    steps: Sequence<usize>,
    reversed: Sequence<usize>,
    #[define(E)] _dtype: StorageType,
) {
    if !value.is_in_bounds(ABSOLUTE_POS) {
        terminate!();
    }

    let rank = comptime![value_shape.len()];
    let mut value_offset = ABSOLUTE_POS;
    let mut input_offset = 0;

    // Calculate the input offset based on value position and slice info
    #[unroll]
    for i in 0..rank {
        // Iterate in reverse to use divmod
        let dim = rank - i - 1;
        let start = starts[dim];
        let end = ends[dim];
        let step = steps[dim];

        let (rem, value_idx) = value_shape[dim].div_mod(value_offset);
        value_offset = rem;

        let input_idx = if reversed[dim] == 0 {
            // Forward stepping
            start + value_idx * step
        } else {
            // Backward stepping - start from end-1
            // For negative steps, we iterate backwards through the selected indices
            let end_minus_1 = end - 1;
            end_minus_1 - value_idx * step
        };

        input_offset += input_idx * input.stride(dim);
    }

    input[input_offset] = value[ABSOLUTE_POS];
}

pub fn slice_assign<R: Runtime>(
    tensor: RudaTensor<R>,
    indices: &[ruda_core::tensor::Slice],
    value: RudaTensor<R>,
) -> RudaTensor<R> {
    assert!(indices.iter().all(|slice| slice.step != 0), "Step cannot be zero");
    if value.meta.num_elements() == 0 {
        return tensor;
    }
    // Check if any slice has non-unit step
    let has_non_unit_step = indices.iter().any(|s| s.step != 1);

    if has_non_unit_step {
        // Use slice_assign_with_steps
        return slice_assign_with_steps(tensor, indices, value);
    }

    let client = tensor.client.clone();
    let tensor = match tensor.can_mut() && tensor.is_nonoverlapping() {
        true => tensor,
        false => tensor.copy(),
    };
    let ndims = tensor.meta.num_dims();
    let ranges = (0..ndims)
        .map(|axis| indices.get(axis).copied().unwrap_or_default().to_range(tensor.meta.shape()[axis]))
        .collect::<Vec<_>>();
    let mut slice_shape = tensor.meta.shape().clone();
    for (dim, range) in slice_shape.iter_mut().zip(&ranges) {
        *dim = range.end - range.start;
    }
    let base_offset = ranges.iter().zip(tensor.meta.strides().iter())
        .map(|(range, stride)| range.start * stride).sum::<usize>();
    let vector_size = if ndims > 0 {
        let candidates = client.io_optimized_vector_sizes(tensor.dtype.size()).filter(|&size| {
            base_offset.is_multiple_of(size)
                && tensor_vector_size_parallel(
                    core::iter::once(size), &slice_shape, tensor.meta.strides(), ndims - 1,
                ) == size
        });
        tensor_vector_size_parallel(candidates, value.meta.shape(), value.meta.strides(), ndims - 1)
    } else {
        1
    };

    let mut shape = SequenceArg::<R, FastDivmod<usize>>::new();
    let mut offsets = SequenceArg::<R, usize>::new();

    for range in ranges {
        shape.push(range.end - range.start);
        offsets.push(range.start);
    }

    let working_units = value.meta.num_elements() / vector_size;
    let ruda_dim = RudaDim::new(tensor.client.properties(), working_units);
    let ruda_count = calculate_ruda_count_elemwise(&tensor.client, working_units, ruda_dim);

    unsafe {
        slice_assign_kernel::launch_unchecked(
            &tensor.client,
            ruda_count,
            ruda_dim,
            address_type!(tensor, value),
            vector_size,
            tensor.clone().into_tensor_arg(),
            value.into_linear_view(),
            shape,
            offsets,
            tensor.dtype.into(),
        )
    };

    tensor
}

/// Slice assign with steps support
///
/// This function handles slice assignment with arbitrary step values, including negative steps.
/// It follows NumPy/PyTorch semantics where values[i] is assigned to selected_indices[i].
///
/// For example, with s![0..6;-1] which selects indices [5,4,3,2,1,0]:
/// - values[0] goes to index 5
/// - values[1] goes to index 4
/// - etc.
pub fn slice_assign_with_steps<R: Runtime>(
    tensor: RudaTensor<R>,
    slices: &[ruda_core::tensor::Slice],
    value: RudaTensor<R>,
) -> RudaTensor<R> {
    assert!(slices.iter().all(|slice| slice.step != 0), "Step cannot be zero");
    if value.meta.num_elements() == 0 {
        return tensor;
    }
    let tensor = match tensor.can_mut() && tensor.is_nonoverlapping() {
        true => tensor,
        false => tensor.copy(),
    };

    // Prepare sequences for kernel
    let mut starts = SequenceArg::<R, usize>::new();
    let mut ends = SequenceArg::<R, usize>::new();
    let mut steps = SequenceArg::<R, usize>::new();
    let mut reversed = SequenceArg::<R, usize>::new();

    for (dim, slice) in slices.iter().enumerate() {
        let range = slice.to_range(tensor.meta.shape()[dim]);
        starts.push(range.start);
        ends.push(range.end);
        steps.push(slice.step.unsigned_abs());
        reversed.push(usize::from(slice.is_reversed()));
    }

    // Pad with default values if needed to match tensor dimensions
    for dim in slices.len()..tensor.meta.num_dims() {
        starts.push(0);
        ends.push(tensor.meta.shape[dim]);
        steps.push(1);
        reversed.push(0);
    }

    // Launch kernel
    let working_units = value.meta.num_elements();
    let ruda_dim = RudaDim::new(tensor.client.properties(), working_units);
    let ruda_count = calculate_ruda_count_elemwise(&tensor.client, working_units, ruda_dim);

    let shape = shape_divmod(&value);
    let step_address = AddressType::from_len(
        slices.iter().map(|slice| slice.step.unsigned_abs()).max().unwrap_or(1),
    );
    unsafe {
        slice_assign_with_steps_kernel::launch_unchecked(
            &tensor.client,
            ruda_count,
            ruda_dim,
            address_type!(tensor, value).max(step_address),
            tensor.clone().into_tensor_arg(),
            value.into_linear_view(),
            shape,
            starts,
            ends,
            steps,
            reversed,
            tensor.dtype.into(),
        );
    }

    tensor
}

/// Helper function for unwrap
#[allow(unused)]
#[ruda]
fn unwrap(value: u32) -> comptime_type!(u32) {
    intrinsic!(|_| value.constant().unwrap().as_u32())
}