use crate::error::Result;
use crate::tensor::Tensor;
use wgpu::util::DeviceExt;
pub struct ReplicationPad2D {
input: Tensor,
padding: (usize, usize, usize, usize), }
impl ReplicationPad2D {
pub fn new(input: Tensor, padding: (usize, usize, usize, usize)) -> Result<Self> {
let shape = input.shape();
if shape.len() != 4 {
return Err(crate::error::BarracudaError::invalid_op(
"ReplicationPad2D",
format!("Expected 4D tensor (NCHW), got {}D", shape.len()),
));
}
Ok(Self { input, padding })
}
fn wgsl_shader() -> &'static str {
{
static S: std::sync::LazyLock<String> = std::sync::LazyLock::new(|| {
include_str!(
"../shaders/tensor/replication_pad_f64.wgsl"
).to_string()
});
&S
}
}
pub fn execute(self) -> Result<Tensor> {
let device = self.input.device();
let shape = self.input.shape();
let batch = shape[0];
let channels = shape[1];
let in_height = shape[2];
let in_width = shape[3];
let (pad_left, pad_right, pad_top, pad_bottom) = self.padding;
let out_height = in_height + pad_top + pad_bottom;
let out_width = in_width + pad_left + pad_right;
let output_size = batch * channels * out_height * out_width;
let input_buffer = self.input.buffer();
let output_buffer = device.device.create_buffer(&wgpu::BufferDescriptor {
label: Some("ReplicationPad2D Output"),
size: (output_size * std::mem::size_of::<f32>()) as u64,
usage: wgpu::BufferUsages::STORAGE | wgpu::BufferUsages::COPY_SRC,
mapped_at_creation: false,
});
#[repr(C)]
#[derive(Copy, Clone, bytemuck::Pod, bytemuck::Zeroable)]
struct Params {
batch: u32,
channels: u32,
in_height: u32,
in_width: u32,
out_height: u32,
out_width: u32,
pad_top: u32,
pad_left: u32,
}
let params = Params {
batch: batch as u32,
channels: channels as u32,
in_height: in_height as u32,
in_width: in_width as u32,
out_height: out_height as u32,
out_width: out_width as u32,
pad_top: pad_top as u32,
pad_left: pad_left as u32,
};
let params_buffer = device.device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
label: Some("ReplicationPad2D Params"),
contents: bytemuck::cast_slice(&[params]),
usage: wgpu::BufferUsages::UNIFORM | wgpu::BufferUsages::COPY_DST,
});
let bind_group_layout = device.device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
label: Some("ReplicationPad2D Bind Group Layout"),
entries: &[
wgpu::BindGroupLayoutEntry {
binding: 0,
visibility: wgpu::ShaderStages::COMPUTE,
ty: wgpu::BindingType::Buffer {
ty: wgpu::BufferBindingType::Storage { read_only: true },
has_dynamic_offset: false,
min_binding_size: None,
},
count: None,
},
wgpu::BindGroupLayoutEntry {
binding: 1,
visibility: wgpu::ShaderStages::COMPUTE,
ty: wgpu::BindingType::Buffer {
ty: wgpu::BufferBindingType::Storage { read_only: false },
has_dynamic_offset: false,
min_binding_size: None,
},
count: None,
},
wgpu::BindGroupLayoutEntry {
binding: 2,
visibility: wgpu::ShaderStages::COMPUTE,
ty: wgpu::BindingType::Buffer {
ty: wgpu::BufferBindingType::Uniform,
has_dynamic_offset: false,
min_binding_size: None,
},
count: None,
},
],
});
let bind_group = device.device.create_bind_group(&wgpu::BindGroupDescriptor {
label: Some("ReplicationPad2D Bind Group"),
layout: &bind_group_layout,
entries: &[
wgpu::BindGroupEntry {
binding: 0,
resource: input_buffer.as_entire_binding(),
},
wgpu::BindGroupEntry {
binding: 1,
resource: output_buffer.as_entire_binding(),
},
wgpu::BindGroupEntry {
binding: 2,
resource: params_buffer.as_entire_binding(),
},
],
});
let shader_module = device.compile_shader(Self::wgsl_shader(), Some("ReplicationPad2D Shader"));
let pipeline_layout = device.device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
label: Some("ReplicationPad2D Pipeline Layout"),
bind_group_layouts: &[&bind_group_layout],
immediate_size: 0,
});
let compute_pipeline = device.device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {
label: Some("ReplicationPad2D Pipeline"),
layout: Some(&pipeline_layout),
module: &shader_module,
entry_point: Some("main"),
cache: None,
compilation_options: Default::default(),
});
let mut encoder = device.create_encoder_guarded(&wgpu::CommandEncoderDescriptor {
label: Some("ReplicationPad2D Encoder"),
});
{
let mut compute_pass = encoder.begin_compute_pass(&wgpu::ComputePassDescriptor {
label: Some("ReplicationPad2D Pass"),
timestamp_writes: None,
});
compute_pass.set_pipeline(&compute_pipeline);
compute_pass.set_bind_group(0, Some(&bind_group), &[]);
let workgroups_x = (out_width as u32 + 15) / 16;
let workgroups_y = (out_height as u32 + 15) / 16;
compute_pass.dispatch_workgroups(workgroups_x, workgroups_y, 1);
}
device.submit_commands(Some(encoder.finish()));
let output_data = crate::utils::read_buffer(device, &output_buffer, output_size)?;
Ok(Tensor::new(
output_data,
vec![batch, channels, out_height, out_width],
device.clone(),
))
}
}
impl Tensor {
pub fn replication_pad2d(self, padding: (usize, usize, usize, usize)) -> Result<Self> {
ReplicationPad2D::new(self, padding)?.execute()
}
}
#[cfg(test)]
mod tests {
use super::*;
#[tokio::test]
async fn test_replication_pad2d_basic() {
let device = crate::device::test_pool::get_test_device().await;
let input = Tensor::new(vec![1.0, 2.0, 3.0, 4.0], vec![1, 1, 2, 2], device.clone());
let output = input.replication_pad2d((1, 1, 1, 1)).unwrap();
assert_eq!(output.shape(), &[1, 1, 4, 4]);
}
#[tokio::test]
async fn test_replication_pad2d_edge_cases() {
let device = crate::device::test_pool::get_test_device().await;
let input = Tensor::new(vec![1.0; 1 * 1 * 4 * 4], vec![1, 1, 4, 4], device.clone());
let output = input.replication_pad2d((0, 0, 0, 0)).unwrap();
assert_eq!(output.shape(), &[1, 1, 4, 4]);
let input = Tensor::new(vec![5.0], vec![1, 1, 1, 1], device.clone());
let output = input.replication_pad2d((2, 2, 2, 2)).unwrap();
assert_eq!(output.shape(), &[1, 1, 5, 5]);
}
#[tokio::test]
async fn test_replication_pad2d_boundary() {
let device = crate::device::test_pool::get_test_device().await;
let input = Tensor::new(vec![1.0; 1 * 3 * 4 * 4], vec![1, 3, 4, 4], device.clone());
let output = input.replication_pad2d((3, 3, 3, 3)).unwrap();
assert_eq!(output.shape(), &[1, 3, 10, 10]);
let input = Tensor::new(vec![1.0; 1 * 1 * 8 * 8], vec![1, 1, 8, 8], device.clone());
let output = input.replication_pad2d((2, 3, 1, 4)).unwrap();
assert_eq!(output.shape(), &[1, 1, 13, 13]);
}
#[tokio::test]
async fn test_replication_pad2d_large_batch() {
let device = crate::device::test_pool::get_test_device().await;
let batch_size = 4;
let channels = 3;
let input = Tensor::new(
vec![1.0; batch_size * channels * 16 * 16],
vec![batch_size, channels, 16, 16],
device.clone(),
);
let output = input.replication_pad2d((2, 2, 2, 2)).unwrap();
assert_eq!(output.shape(), &[batch_size, channels, 20, 20]);
}
}