forked from pytorch/pytorch
-
Notifications
You must be signed in to change notification settings - Fork 0
/
THCScanUtils.cuh
218 lines (185 loc) · 6.51 KB
/
THCScanUtils.cuh
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
#ifndef THC_SCAN_UTILS_INC
#define THC_SCAN_UTILS_INC
#include <THC/THCAsmUtils.cuh>
#include <THC/THCDeviceUtils.cuh>
#include <c10/macros/Macros.h>
// Collection of in-kernel scan / prefix sum utilities
// Inclusive Scan via an upsweep/downsweep mechanism. Assumes:
//
// 1. Power2ScanSize is a power of 2. This code still works for collections that
// do not exactly contain a power of 2 number of elements, simply round up to the
// nearest power of 2 and then call.
//
// 2. That there are two-elements per thread, i.e. the size of the smem storage
// is 2 * blockDim.x * sizeof(T).
//
// Consider a (+)-Scan on the following elements:
//
// Upsweep:
//
// 0 1 2 3 4 5 6 7
// 1 5 9 13
// 6 22
// 28
//
// Downsweep:
// 15
// 3 10 21
template <typename T, class BinaryOp, int Power2ScanSize>
__device__ void inclusivePrefixScan(T *smem, BinaryOp binop) {
// Reduce step ("upsweep")
#pragma unroll
for (int stride = 1; stride < Power2ScanSize; stride <<= 1) {
int index = (threadIdx.x + 1) * stride * 2 - 1;
if (index < Power2ScanSize) {
smem[index] = binop(smem[index], smem[index - stride]);
}
__syncthreads();
}
// Post-reduce step ("downsweep")
#pragma unroll
for (int stride = Power2ScanSize / 4; stride > 0; stride >>= 1) {
int index = (threadIdx.x + 1) * stride * 2 - 1;
if ((index + stride) < Power2ScanSize) {
smem[index + stride] = binop(smem[index + stride], smem[index]);
}
__syncthreads();
}
}
// Generic Op that can be used to support segmented scans by re-using
// the basic inclusiveScanOp. Merely requires that the input data has both
// a flag and val component
template <typename T, class BinaryOp>
struct SegmentedScanOp {
__host__ __device__ SegmentedScanOp(BinaryOp binop): _binop(binop) {}
__host__ __device__ inline T operator()(const T& a, const T& b) {
T c;
c.val = a.flag ? a.val : _binop(a.val, b.val);
c.flag = a.flag | b.flag;
return c;
}
BinaryOp _binop;
};
// Extends the above Inclusive Scan to support segments. It has the same properties
// but also takes a flag array that indicates the starts of "segments", i.e. individual
// units to scan. For example, consider the following (+)-scan that is segmented:
//
// Input: [1, 3, 2, 4, 1, 2, 3, 2, 1, 4]
// Flags: [1, 0, 0, 1, 0, 1, 1, 0, 1, 0]
// Output: 1 4 6 4 5 2 3 5 1 5
//
// So we see that each "flag" resets the scan to that index.
template <typename T, class BinaryOp, int Power2ScanSize>
__device__ void segmentedInclusivePrefixScan(T *smem, bool *bmem, BinaryOp binop) {
// Reduce step ("upsweep")
#pragma unroll
for (int stride = 1; stride < Power2ScanSize; stride <<= 1) {
int index = (threadIdx.x + 1) * stride * 2 - 1;
if (index < Power2ScanSize) {
smem[index] = bmem[index] ? smem[index] : binop(smem[index], smem[index - stride]);
bmem[index] = bmem[index] | bmem[index - stride];
}
__syncthreads();
}
// Post-reduce step ("downsweep")
#pragma unroll
for (int stride = Power2ScanSize / 4; stride > 0; stride >>= 1) {
int index = (threadIdx.x + 1) * stride * 2 - 1;
if ((index + stride) < Power2ScanSize) {
smem[index + stride] = bmem[index + stride] ? smem[index + stride] : binop(smem[index + stride], smem[index]);
bmem[index + stride] = bmem[index + stride] | bmem[index];
}
__syncthreads();
}
}
// Inclusive prefix sum using shared memory
template <typename T, bool KillWARDependency, class BinaryFunction>
__device__ void inclusivePrefixScan(T* smem, T in, T* out, BinaryFunction binop) {
// FIXME: this is a slow, simple implementation; need up/down sweep,
// prevent smem conflicts
smem[threadIdx.x] = in;
__syncthreads();
for (int offset = 1; offset < blockDim.x; offset *= 2) {
T val = 0;
if (threadIdx.x >= offset) {
val = binop(smem[threadIdx.x - offset], smem[threadIdx.x]);
}
__syncthreads();
if (threadIdx.x >= offset) {
smem[threadIdx.x] = val;
}
__syncthreads();
}
*out = smem[threadIdx.x];
// Prevent write-after-read dependencies on smem usage above if necessary
if (KillWARDependency) {
__syncthreads();
}
}
// Exclusive prefix sum using shared memory
template <typename T, bool KillWARDependency, class BinaryFunction>
__device__ void exclusivePrefixScan(T* smem, T in, T* out, T* carry, BinaryFunction binop) {
// FIXME: crappy implementation
// We kill write-after-read dependencies separately below, hence the `false`
inclusivePrefixScan<T, false, BinaryFunction>(smem, in, out, binop);
*out -= in;
*carry = smem[blockDim.x - 1];
// Prevent write-after-read dependencies on smem usage above if necessary
if (KillWARDependency) {
__syncthreads();
}
}
// Inclusive prefix sum for binary vars using intra-warp voting +
// shared memory
template <typename T, bool KillWARDependency, class BinaryFunction>
__device__ void inclusiveBinaryPrefixScan(T* smem, bool in, T* out, BinaryFunction binop) {
// Within-warp, we use warp voting.
#if defined (__HIP_PLATFORM_HCC__)
unsigned long long int vote = WARP_BALLOT(in);
T index = __popcll(getLaneMaskLe() & vote);
T carry = __popcll(vote);
#else
T vote = WARP_BALLOT(in);
T index = __popc(getLaneMaskLe() & vote);
T carry = __popc(vote);
#endif
int warp = threadIdx.x / C10_WARP_SIZE;
// Per each warp, write out a value
if (getLaneId() == 0) {
smem[warp] = carry;
}
__syncthreads();
// Sum across warps in one thread. This appears to be faster than a
// warp shuffle scan for CC 3.0+
if (threadIdx.x == 0) {
int current = 0;
for (int i = 0; i < blockDim.x / C10_WARP_SIZE; ++i) {
T v = smem[i];
smem[i] = binop(smem[i], current);
current = binop(current, v);
}
}
__syncthreads();
// load the carry from the preceding warp
if (warp >= 1) {
index = binop(index, smem[warp - 1]);
}
*out = index;
if (KillWARDependency) {
__syncthreads();
}
}
// Exclusive prefix sum for binary vars using intra-warp voting +
// shared memory
template <typename T, bool KillWARDependency, class BinaryFunction>
__device__ void exclusiveBinaryPrefixScan(T* smem, bool in, T* out, T* carry, BinaryFunction binop) {
inclusiveBinaryPrefixScan<T, false, BinaryFunction>(smem, in, out, binop);
// Inclusive to exclusive
*out -= (T) in;
// The outgoing carry for all threads is the last warp's sum
*carry = smem[THCCeilDiv<int>(blockDim.x, C10_WARP_SIZE) - 1];
if (KillWARDependency) {
__syncthreads();
}
}
#endif // THC_SCAN_UTILS_INC