-
Notifications
You must be signed in to change notification settings - Fork 79
/
ipu_options.py
140 lines (114 loc) · 5.09 KB
/
ipu_options.py
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
# Copyright (c) 2021 Graphcore Ltd. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import ctypes
import numpy as np
import poptorch
import popart
import torch
import popdist.poptorch
def create_options(opts):
if opts.use_popdist:
model_opts = popdist.poptorch.Options()
else:
model_opts = poptorch.Options()
return model_opts
def get_options(config):
"""
Set ipu specific options for the model, see documentation:
https://docs.graphcore.ai/en/latest/
"""
# PopTorch options
opts = create_options(config)
opts.autoRoundNumIPUs(True)
opts.deviceIterations(config.device_iterations)
# Set replication and gradient accumulation factors
if not config.use_popdist:
opts.replicationFactor(config.replication_factor)
opts.Training.gradientAccumulation(config.gradient_accumulation)
if config.reduction_type == "sum":
opts.Training.accumulationAndReplicationReductionType(poptorch.ReductionType.Sum)
elif config.reduction_type == "mean":
opts.Training.accumulationAndReplicationReductionType(poptorch.ReductionType.Mean)
else:
raise ValueError("Expected reduction type to be 'sum' or 'mean', but got %s" % config.reduction_type)
# Enable automatic loss scaling
# Note that it expects accumulationAndReplicationReductionType to be set
# to Mean as above, and for accumulation by the optimizer to be done in
# half precision using accum_type=torch.float16 during optimizer instantiation.
if config.auto_loss_scaling is True:
opts.Training.setAutomaticLossScaling(True)
# Return all results from IPU to host
opts.outputMode(poptorch.OutputMode.All)
# Fix the random seeds
np.random.seed(config.random_seed)
opts.randomSeed(config.random_seed)
# Enable Replicated Tensor Sharding (RTS) of optimizer state
# with optimizer state residing either on-chip or in DRAM
opts.TensorLocations.setOptimizerLocation(
poptorch.TensorLocationSettings()
# Optimizer state lives on-chip
.useOnChipStorage(not config.optimizer_state_offchip)
# Shard optimizer state between replicas with zero-redundancy
.useReplicatedTensorSharding(config.enable_rts)
)
# Use Pipelined Execution
opts.setExecutionStrategy(poptorch.PipelinedExecution(poptorch.AutoStage.SameAsIpu))
# Set available Transient Memory For matmuls and convolutions operations
mem_prop = {f"IPU{i}": config.matmul_proportion[i] for i in range(config.ipus_per_replica)}
opts.setAvailableMemoryProportion(mem_prop)
if config.synthetic_data:
opts.enableSyntheticData(int(popart.SyntheticDataMode.RandomNormal))
# Enable stochastic rounding (recommended for training with FP16)
opts.Precision.enableStochasticRounding(config.stochastic_rounding)
# Enable caching the compiled executable to disk
if config.executable_cache_dir:
opts.enableExecutableCaching(config.executable_cache_dir)
# Half precision partials for matmuls and convolutions
if config.half_partials:
opts.Precision.setPartialsType(torch.half)
# PopART performance options
# Only stream needed tensors back to host
opts._Popart.set("disableGradAccumulationTensorStreams", True)
opts._Popart.set(
"accumulateOuterFragmentSettings.schedule", int(popart.AccumulateOuterFragmentSchedule.OverlapMemoryOptimized)
)
if config.prefetch_depth > 1:
# How many batches to prefetch onto the IPU
opts._Popart.set("defaultPrefetchBufferingDepth", config.prefetch_depth)
# Options for profiling with Popvision
engine_options = {
"opt.useAutoloader": "true",
"target.syncReplicasIndependently": "true",
}
if config.profile_dir:
engine_options = {
**engine_options,
**{
"debug.allowOutOfMemory": "true",
"autoReport.directory": config.profile_dir,
"profiler.format": "v3",
"autoReport.all": "true",
},
}
opts._Popart.set("engineOptions", engine_options)
# Parallelize optimizer step update across IPUs
opts._Popart.set(
"accumulateOuterFragmentSettings.schedule", int(popart.AccumulateOuterFragmentSchedule.OverlapMemoryOptimized)
)
opts._Popart.set("accumulateOuterFragmentSettings.excludedVirtualGraphs", ["0"])
# Enable patterns for better throughput and memory reduction
opts._Popart.set("subgraphCopyingStrategy", int(popart.SubgraphCopyingStrategy.JustInTime))
opts._Popart.set("scheduleNonWeightUpdateGradientConsumersEarly", True)
return opts