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

extern crate ocl;

mod kernel;

use ocl::{Platform, Device, ProQue, Buffer};
use ocl::core;
use ocl::core::{DeviceInfo};

pub struct Context{
    pub compute_units: u32,
    pro_que:       ocl::ProQue,
}


pub struct Matrix {
    pub rows: usize,
    pub cols: usize,
    pub data: Vec<f32>,
}

impl Context {
    pub fn mul_matrix_scalar(&mut self, matrix: &Matrix, scalar: f32) -> Matrix {
        let ref mut ocl_pq = self.pro_que;
        
        // set work dimension, we can work with 1
        ocl_pq.set_dims([matrix.rows*matrix.cols]);
        
        // copy matrix to device
        let source_buffer = Buffer::new(
            &ocl_pq.queue().clone(),
            Some(core::MEM_READ_WRITE | core::MEM_COPY_HOST_PTR), 
            ocl_pq.dims().clone(), 
            Some(&matrix.data)).unwrap();
        
        // prepare result matrix
        let mut result = vec![0.0f32; matrix.cols*matrix.rows];
        let result_buffer: Buffer<f32> = ocl_pq.create_buffer().unwrap();

        // Create a kernel with arguments corresponding to those in the kernel:
        let kernel = ocl_pq.create_kernel("mul_matrix_scalar").unwrap()
            .arg_scl(scalar)
            .arg_buf(&source_buffer)
            .arg_buf(&result_buffer);

        // Enqueue kernel:
        kernel.enq().unwrap(); // send kernel to device and run it

        // Read results from the device into result_buffer's local vector:
        result_buffer.read(&mut result).enq().unwrap();
        
        // return matrix
        Matrix{rows: matrix.rows,cols: matrix.cols, data: result}
    }
    
    pub fn mul_matrix_matrix(&mut self, matrix_a: &Matrix, matrix_b: &Matrix) -> Matrix {
        let ref mut ocl_pq = self.pro_que;
        
        // set work dimension for matrix_a
        ocl_pq.set_dims([matrix_a.rows,matrix_a.cols]);
        
        // copy matrix_a to device
        let matrix_a_buffer = Buffer::new(
            &ocl_pq.queue().clone(),
            Some(core::MEM_READ_WRITE | core::MEM_COPY_HOST_PTR), 
            ocl_pq.dims().clone(), 
            Some(&matrix_a.data)).unwrap();

        // set work dimension for matrix_b
        ocl_pq.set_dims([matrix_b.rows,matrix_b.cols]);
        // copy matrix_b to device
        let matrix_b_buffer = Buffer::new(
            &ocl_pq.queue().clone(),
            Some(core::MEM_READ_WRITE | core::MEM_COPY_HOST_PTR), 
            ocl_pq.dims().clone(), 
            Some(&matrix_b.data)).unwrap();
        
        // set work dimension for result matrix
        ocl_pq.set_dims([matrix_a.rows,matrix_b.cols]);
        // prepare result matrix
        
        let mut result = vec![0.0f32; matrix_a.rows*matrix_b.cols];
        let result_buffer: Buffer<f32> = ocl_pq.create_buffer().unwrap();

        // Create a kernel with arguments corresponding to those in the kernel:
        let kernel = ocl_pq.create_kernel("mul_matrix_matrix").unwrap()
            .arg_buf(&matrix_a_buffer)
            .arg_buf(&matrix_b_buffer)
            .arg_buf(&result_buffer)
            .arg_scl(matrix_a.rows as i32)
            .arg_scl(matrix_a.cols as i32)
            .arg_scl(matrix_b.cols as i32);
        //panic!("{}",kernel.get_lws().to_len());

        // Enqueue kernel:
        kernel.enq().unwrap(); // send kernel to device and run it

        // Read results from the device into result_buffer's local vector:
        result_buffer.read(&mut result).enq().unwrap();
        
        // return matrix
        Matrix{rows: matrix_a.rows, cols: matrix_b.cols, data: result}
    }
}

pub fn new() -> Option<Context> {
    // wich Device should we choose?
    // the one with the most Compute units!
    let mut compute_units = 0;
    let mut ocl_device = None;
    let platforms = Platform::list();
    for p_idx in 0..platforms.len() {
        let platform = &platforms[p_idx];
        let devices = Device::list_all(platform);
        for d_idx in 0..devices.len() {
            let device = devices[d_idx];
            let deviceinforesult = core::get_device_info(&device, DeviceInfo::MaxComputeUnits);
            let units = deviceinforesult.to_string().parse().unwrap();
            if units > compute_units {
                ocl_device = Some(device);
                compute_units = units;
            }
        }
    }
    // something went wrong no, opencl not installed
    if compute_units == 0 {
        return None
    }
    let que = ProQue::builder()
              .device(ocl_device.unwrap())
              .src(kernel::OCL_KERNEL)
              .build().expect("Build ProQue");
    Some(Context{
        compute_units: compute_units,
        pro_que : que,
    })
}

#[test]
fn single_test() {
    let eps = 1.0e-6;
    let mut c = new().unwrap();
    let m0 = Matrix{rows: 2, cols: 2, data: vec![1.0, 2.0, 3.0, 4.0]};
    let m1 = Matrix{rows: 2, cols: 1, data: vec![4.0, 5.0]};
    
    /* matrix multiplication
       [1 2] * [4] = [14]
       [3 4]   [5]   [32] */
    let m = c.mul_matrix_matrix(&m0,&m1);
    assert!((m.data[0] - 14.0f32).abs() < eps);
    assert!((m.data[1] - 32.0f32).abs() < eps);
    
    /* matrix scalar multiplication 
       [1 2] * 1.5 = [1.5 3]
       [3 4]         [4.5 6]*/
    let m = c.mul_matrix_scalar(&m0,1.5);
    assert!( (m.data[0] - 1.5f32) < eps);
    assert!( (m.data[1] - 3.0f32) < eps);
    assert!( (m.data[2] - 4.5f32) < eps);
    assert!( (m.data[3] - 6.0f32) < eps);
}