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
|
// BSD 3-Clause License; see https://github.com/scikit-hep/awkward-1.0/blob/main/LICENSE
#define FILENAME(line) FILENAME_FOR_EXCEPTIONS_CUDA("src/cpu-kernels/awkward_IndexedArray_getitem_nextcarry.cu", line)
#include "standard_parallel_algorithms.h"
#include "awkward/kernels.h"
template <typename C, typename T>
__global__
void awkward_IndexedArray_getitem_nextcarry_kernel(
T* tocarry,
const C* fromindex,
int64_t* prefixedsum_mask,
int64_t lenindex,
int64_t lencontent,
unsigned long long* error_i) {
/**
* Here the thread_id has a unsigned long long data type rather than a int64_t
* type because atomicMin doesn't provide a fucntion signature for int64_t type
*/
unsigned long long thread_id = blockIdx.x * blockDim.x + threadIdx.x;
if (thread_id < lenindex) {
C j = fromindex[thread_id];
if (j < 0 || j >= lencontent) {
atomicMin(error_i, thread_id);
} else {
tocarry[prefixedsum_mask[thread_id] - 1] = j;
}
}
}
template <typename C>
__global__ void
awkward_IndexedArray_getitem_nextcarry_filter_mask(
const C* fromindex,
int8_t* filtered_mask,
int64_t lenindex,
int64_t lencontent) {
int64_t thread_id = blockIdx.x * blockDim.x + threadIdx.x;
if (thread_id < lenindex) {
if (!(fromindex[thread_id] < 0 || fromindex[thread_id] >= lencontent)) {
{
filtered_mask[thread_id] = 1;
}
}
}
}
__global__ void
awkward_IndexedArray_getitem_nextcarry_initialize_error_i(
unsigned long long* error_i,
unsigned long long value) {
*error_i = value;
}
template <typename C, typename T>
ERROR awkward_IndexedArray_getitem_nextcarry(
T* tocarry,
const C* fromindex,
int64_t lenindex,
int64_t lencontent) {
dim3 blocks_per_grid = blocks(lenindex);
dim3 threads_per_block = threads(lenindex);
int8_t* filtered_mask;
int64_t* res_temp;
HANDLE_ERROR(cudaMalloc((void**)&filtered_mask, sizeof(int8_t) * lenindex));
HANDLE_ERROR(cudaMalloc((void**)&res_temp, sizeof(int64_t) * lenindex));
HANDLE_ERROR(cudaMemset(filtered_mask, 0, sizeof(int8_t) * lenindex));
awkward_IndexedArray_getitem_nextcarry_filter_mask<C><<<blocks_per_grid, threads_per_block>>>(
fromindex,
filtered_mask,
lenindex,
lencontent);
exclusive_scan(res_temp, filtered_mask, lenindex);
unsigned long long * dev_error_i;
unsigned long long error_i;
HANDLE_ERROR(cudaMalloc((void**)&dev_error_i, sizeof(unsigned long long)));
awkward_IndexedArray_getitem_nextcarry_initialize_error_i<<<1,1>>>(
dev_error_i,
lenindex + 1);
awkward_IndexedArray_getitem_nextcarry_kernel<C, T><<<blocks_per_grid, threads_per_block>>>(
tocarry,
fromindex,
res_temp,
lenindex,
lencontent,
dev_error_i);
HANDLE_ERROR(cudaMemcpy(&error_i, dev_error_i, sizeof(unsigned long long), cudaMemcpyDeviceToHost));
if(error_i != lenindex + 1) {
C error_j;
HANDLE_ERROR(cudaMemcpy(&error_j, fromindex + error_i, sizeof(C), cudaMemcpyDeviceToHost));
return failure("index out of range", error_i, error_j, FILENAME(__LINE__));
}
return success();
}
ERROR awkward_IndexedArray32_getitem_nextcarry_64(
int64_t* tocarry,
const int32_t* fromindex,
int64_t lenindex,
int64_t lencontent) {
return awkward_IndexedArray_getitem_nextcarry<int32_t, int64_t>(
tocarry,
fromindex,
lenindex,
lencontent);
}
ERROR awkward_IndexedArrayU32_getitem_nextcarry_64(
int64_t* tocarry,
const uint32_t* fromindex,
int64_t lenindex,
int64_t lencontent) {
return awkward_IndexedArray_getitem_nextcarry<uint32_t, int64_t>(
tocarry,
fromindex,
lenindex,
lencontent);
}
ERROR awkward_IndexedArray64_getitem_nextcarry_64(
int64_t* tocarry,
const int64_t* fromindex,
int64_t lenindex,
int64_t lencontent) {
return awkward_IndexedArray_getitem_nextcarry<int64_t, int64_t>(
tocarry,
fromindex,
lenindex,
lencontent);
}
|
