cera 0.5.0

Rust-native LLM inference engine
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

#include <metal_stdlib>
#include <metal_math>
#include <metal_texture>
using namespace metal;

#line 49 "cera/src/backend/shaders/slang/conv1d.slang"
struct KernelContext_0
{
    uint device* par_buf_0;
    float device* rbuf_0;
    float device* w_buf_0;
    float device* in_buf_0;
    float device* out_buf_0;
};


#line 36
[[kernel]] void conv1d_depthwise(uint3 gid_0 [[thread_position_in_grid]], uint device* par_buf_1 [[buffer(4)]], float device* rbuf_1 [[buffer(1)]], float device* w_buf_1 [[buffer(2)]], float device* in_buf_1 [[buffer(0)]], float device* out_buf_1 [[buffer(3)]])
{

#line 36
    thread KernelContext_0 kernelContext_0;

#line 36
    (&kernelContext_0)->par_buf_0 = par_buf_1;

#line 36
    (&kernelContext_0)->rbuf_0 = rbuf_1;

#line 36
    (&kernelContext_0)->w_buf_0 = w_buf_1;

#line 36
    (&kernelContext_0)->in_buf_0 = in_buf_1;

#line 36
    (&kernelContext_0)->out_buf_0 = out_buf_1;
    uint ch_0 = gid_0.x;
    uint hs_0 = par_buf_1[int(0)];
    uint ks_0 = par_buf_1[int(1)];
    uint d_conv_0 = par_buf_1[int(2)];

    if(ch_0 >= hs_0)
    {

#line 43
        return;
    }

#line 43
    uint k_0 = 0U;

#line 43
    float sum_0 = 0.0f;

#line 48
    for(;;)
    {

#line 48
        if(k_0 < d_conv_0)
        {
        }
        else
        {

#line 48
            break;
        }

#line 49
        float sum_1 = sum_0 + *((&kernelContext_0)->rbuf_0+(k_0 * hs_0 + ch_0)) * (&kernelContext_0)->w_buf_0[ch_0 * ks_0 + k_0];

#line 48
        k_0 = k_0 + 1U;

#line 48
        sum_0 = sum_1;

#line 48
    }



    *((&kernelContext_0)->out_buf_0+ch_0) = sum_0 + (&kernelContext_0)->in_buf_0[ch_0] * (&kernelContext_0)->w_buf_0[ch_0 * ks_0 + d_conv_0];

#line 57
    if(d_conv_0 > 1U)
    {

#line 57
        k_0 = 0U;
        for(;;)
        {

#line 58
            if(k_0 < (d_conv_0 - 1U))
            {
            }
            else
            {

#line 58
                break;
            }

#line 59
            uint _S1 = k_0 + 1U;

#line 59
            *((&kernelContext_0)->rbuf_0+(k_0 * hs_0 + ch_0)) = *((&kernelContext_0)->rbuf_0+(_S1 * hs_0 + ch_0));

#line 58
            k_0 = _S1;

#line 58
        }

#line 57
    }

#line 62
    if(d_conv_0 > 0U)
    {

#line 63
        *((&kernelContext_0)->rbuf_0+((d_conv_0 - 1U) * hs_0 + ch_0)) = (&kernelContext_0)->in_buf_0[ch_0];

#line 62
    }


    return;
}