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
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
|
## @package muji
# Module caffe2.python.muji
"""muji.py does multi-gpu training for caffe2 with no need to change the c++
side code. Everything is defined on the computation graph level.
We support the following use cases:
- 2 gpus, where peer access is enabled between them.
- 4 gpus, where peer access are enabled between all of them.
- 4 gpus, where peer access are enabled in two groups,
between {1, 2} and {3, 4}
- 8 gpus, where peer access are enabled in two groups,
between {1, 2, 3, 4} and {5, 6, 7, 8}.
If above cases are not satisfied, a fallback function which does not rely on
peer access will be called.
"""
import numpy as np
from caffe2.proto import caffe2_pb2
from caffe2.python import workspace
def OnGPU(gpu_id):
"""A utility function that returns a device option protobuf of the
specified gpu id.
"""
device_option = caffe2_pb2.DeviceOption()
device_option.device_type = workspace.GpuDeviceType
device_option.device_id = gpu_id
return device_option
def OnCPU():
device_option = caffe2_pb2.DeviceOption()
device_option.device_type = caffe2_pb2.CPU
return device_option
def Allreduce(net, blobs, reduced_affix="_reduced", gpu_indices=None):
"""The general Allreduce interface that reroutes the function calls.
CPUs and AMD GPUs are not supported because
GetGpuPeerAccessPattern is called to get gpu peer access pattern.
"""
if gpu_indices is None:
gpu_indices = list(range(len(blobs)))
if len(gpu_indices) != len(blobs):
raise RuntimeError(
"gpu_indices length and blobs length mismatch: %d vs %d" %
(len(gpu_indices), len(blobs))
)
pattern = workspace.GetGpuPeerAccessPattern()
if len(blobs) == 2 and pattern.shape[0] >= 2 and np.all(pattern[:2, :2]):
return Allreduce2(net, blobs, reduced_affix, gpu_indices)
elif len(blobs) == 4 and pattern.shape[0] >= 4 and np.all(pattern[:4, :4]):
return Allreduce4(net, blobs, reduced_affix, gpu_indices)
elif len(blobs) == 4 and pattern.shape[0] >= 4 and np.all(pattern[:2, :2]) and np.all(pattern[2:4, 2:4]):
return Allreduce4Group2(net, blobs, reduced_affix, gpu_indices)
elif len(blobs) == 8 and pattern.shape[0] >= 8 and np.all(pattern[:8, :8]):
return Allreduce8(net, blobs, reduced_affix, gpu_indices)
else:
return AllreduceFallback(net, blobs, reduced_affix, gpu_indices)
def Allreduce2(net, blobs, reduced_affix, gpu_indices):
"""Allreduce for 2 gpus.
Algorithm: 0r <- 0 + 1, 1r <- 0r, where r means "reduced"
"""
a, b = blobs
gpu_a, gpu_b = gpu_indices
a_reduced = net.Add([a, b], a + reduced_affix, device_option=OnGPU(gpu_a))
b_reduced = a_reduced.Copy(
[],
b + reduced_affix,
device_option=OnGPU(gpu_b)
)
return a_reduced, b_reduced
def Allreduce4(net, blobs, reduced_affix, gpu_indices):
"""Allreduce for 4 gpus.
Algorithm: 2 level reduction.
0r <- 0 + 1, 2r <- 2 + 3
0r <- 0r + 2r
2r <- 0r,
1r <- 0r, 3r <- 2r
"""
a, b, c, d = blobs
gpu_a, gpu_b, gpu_c, gpu_d = gpu_indices
# a_reduced <- a+b, c_reduced <- c + d
a_reduced = net.Add(
[a, b],
str(a) + reduced_affix,
device_option=OnGPU(gpu_a)
)
c_reduced = net.Add(
[c, d],
str(c) + reduced_affix,
device_option=OnGPU(gpu_c)
)
# a_reduced <- a_reduced + c_reduced
a_reduced = a_reduced.Add(c_reduced, a_reduced, device_option=OnGPU(gpu_a))
# broadcast a_reduced to c_reduced
c_reduced = a_reduced.Copy([], c_reduced, device_option=OnGPU(gpu_c))
# broadcast to b and d
b_reduced = a_reduced.Copy(
[],
str(b) + reduced_affix,
device_option=OnGPU(gpu_b)
)
d_reduced = c_reduced.Copy(
[],
str(d) + reduced_affix,
device_option=OnGPU(gpu_d)
)
return a_reduced, b_reduced, c_reduced, d_reduced
def Allreduce4Group2(net, blobs, reduced_affix, gpu_indices):
"""Allreduce for 4 gpus where peer access are enabled in {0,1} and {2,3}
Algorithm: 2 level reduction.
0r <- 0 + 1, 2r <- 2 + 3
0r <- 0r + 2r
2r <- 0r,
1r <- 0r, 3r <- 2r
"""
a, b, c, d = blobs
gpu_a, gpu_b, gpu_c, gpu_d = gpu_indices
# a_reduced <- a+b, c_reduced <- c + d
a_reduced = net.Add(
[a, b],
str(a) + reduced_affix,
device_option=OnGPU(gpu_a)
)
c_reduced = net.Add(
[c, d],
str(c) + reduced_affix,
device_option=OnGPU(gpu_c)
)
# copy from c_reduce(gpu_c) to c_reduce_copy(gpu_a)
c_reduced_copy = c_reduced.Copy(
[],
str(c_reduced) + '_copy',
device_option=OnGPU(gpu_a)
)
# a_reduced <- a_reduced + c_reduced_copy
a_reduced = a_reduced.Add(c_reduced_copy, a_reduced, device_option=OnGPU(gpu_a))
# broadcast a_reduced to c_reduced
c_reduced = a_reduced.Copy([], c_reduced, device_option=OnGPU(gpu_c))
# broadcast to b and d
b_reduced = a_reduced.Copy(
[],
str(b) + reduced_affix,
device_option=OnGPU(gpu_b)
)
d_reduced = c_reduced.Copy(
[],
str(d) + reduced_affix,
device_option=OnGPU(gpu_d)
)
return a_reduced, b_reduced, c_reduced, d_reduced
def Allreduce8(net, blobs, reduced_affix, gpu_indices):
"""Allreduce for 8 gpus.
Algorithm: 3 level reduction.
0r <- 0 + 1, 2r <- 2 + 3, 4r <- 4 + 5, 6r <- 6 + 7
0r <- 0r + 2r, 4r <- 4r + 6r
0r <- 0r + 4r
4r <- 0r
2r <- 0r, 6r <- 4r
1r <- 0r, 3r <- 2r, 5r <- 4r, 7r <- 6r
"""
reduced = [None] * 8
# Reduction level 1
for i in [0, 2, 4, 6]:
reduced[i] = net.Add(
[blobs[i], blobs[i + 1]],
blobs[i] + reduced_affix,
device_option=OnGPU(gpu_indices[i])
)
# Reduction level 2
for i in [0, 4]:
reduced[i] = net.Add(
[reduced[i], reduced[i + 2]],
str(blobs[i]) + reduced_affix,
device_option=OnGPU(gpu_indices[i])
)
# Reduction level 3: this involves a copy.
reduced_4_copy = reduced[4].Copy(
[],
str(reduced[4]) + '_copy',
device_option=OnGPU(gpu_indices[0])
)
reduced[0] = reduced[0].Add(
reduced_4_copy,
reduced[0],
device_option=OnGPU(gpu_indices[0])
)
# Broadcast level 1
reduced[4] = reduced[0].Copy(
[],
reduced[4],
device_option=OnGPU(gpu_indices[4])
)
# Broadcast level 2
for i in [2, 6]:
reduced[i] = reduced[i - 2].Copy(
[],
reduced[i],
device_option=OnGPU(gpu_indices[i])
)
# Broadcast level 3
for i in [1, 3, 5, 7]:
reduced[i] = reduced[i - 1].Copy(
[],
blobs[i] + reduced_affix,
device_option=OnGPU(gpu_indices[i])
)
return reduced
def AllreduceFallback(net, blobs, reduced_affix, gpu_indices):
"""A fallback option for Allreduce with no assumption on p2p.
Algorithm: a flat operation on gpu 0
0r <- 0
0r <- 0r + i for i in gpu_indices[1:]
ir <- 0r for i in gpu_indices[1:]
"""
reduced = [None] * len(gpu_indices)
if reduced_affix != '':
# copy first
reduced[0] = net.Copy(
blobs[0],
blobs[0] + reduced_affix,
device_option=OnGPU(gpu_indices[0])
)
else:
reduced[0] = blobs[0]
# do temp copy and add
temp_name = reduced[0] + '_temp_copy'
for i in range(1, len(gpu_indices)):
temp = net.Copy(
blobs[i],
temp_name,
device_option=OnGPU(gpu_indices[0])
)
reduced[0] = net.Add(
[temp, reduced[0]],
reduced[0],
device_option=OnGPU(gpu_indices[0])
)
# Broadcast to everyone else
for i in range(1, len(gpu_indices)):
reduced[i] = net.Copy(
reduced[0],
blobs[i] + reduced_affix,
device_option=OnGPU(gpu_indices[i])
)
return reduced
|
