Tensorflow: MultiWorkerMirroredStrategy 성능이 낮음 (2gpu, 2node) X1.3 속도 향상

시스템 정보

Have I written custom code (as opposed to using a stock example script provided in TensorFlow):
OS Platform and Distribution: Ubuntu 18.04
TensorFlow installed from (source or binary): pip install tensorflow-gpu
TensorFlow version (use command below): 2.0
Python version: 3.6.9
CUDA/cuDNN version: 10/7.6.4.38
GPU model and memory: Tesla P4  8G

현재 행동 설명
아래에 설명 된 코드를 실행합니다.

테스트 1 : (2 대의 기계)

os.environ [ 'TF_CONFIG'] = json.dumps ({
'클러스터': {
'작업자': [ "server1 : 12345", "server2 : 12345"]
},
'작업': { '유형': '작업자', '색인': 0}
})

다른 기계에서

os.environ [ 'TF_CONFIG'] = json.dumps ({
'클러스터': {
'작업자': [ "server1 : 12345", "server2 : 12345"]
},
'작업': { '유형': '작업자', '색인': 1}
})

스크립트가 첫 번째 에포크를 처리하기 시작하면 충돌합니다.

예상되는 동작 설명

15 초 / epoch가 너무 느립니다.

테스트 2 : (한 대의 컴퓨터)

os.environ [ 'TF_CONFIG'] = json.dumps ({
'클러스터': {
'작업자': [ "server1 : 12345"]
},
'작업': { '유형': '작업자', '색인': 0}
})

예상되는 동작 설명

사용 전략과 동일한 5s / epoch = tf.distribute.MirroredStrategy () GPU 카드 1 개

암호

import ssl
import os
import json
import argparse
import time

import numpy as np
import tensorflow as tf
import tensorflow_datasets as tfds

ssl._create_default_https_context = ssl._create_unverified_context


def configure_cluster(worker_hosts=None, task_index=-1):
    """Set multi-worker cluster spec in TF_CONFIG environment variable.
    Args:
      worker_hosts: comma-separated list of worker ip:port pairs.
    Returns:
      Number of workers in the cluster.
    """
    tf_config = json.loads(os.environ.get('TF_CONFIG', '{}'))
    if tf_config:
        num_workers = len(tf_config['cluster'].get('worker', []))
    elif worker_hosts:
        workers = worker_hosts.split(',')
        num_workers = len(workers)
        if num_workers > 1 and task_index < 0:
            raise ValueError('Must specify task_index when number of workers > 1')
        task_index = 0 if num_workers == 1 else task_index
        os.environ['TF_CONFIG'] = json.dumps({
            'cluster': {
                'worker': workers
            },
            'task': {'type': 'worker', 'index': task_index}
        })
    else:
        num_workers = 1
    return num_workers


parser = argparse.ArgumentParser(description='TensorFlow Benchmark',
                                 formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--num-epochs', type=int, default=5, help='input batch size')
parser.add_argument('--batch-size-per-replica', type=int, default=32, help='input batch size')
parser.add_argument('--worker-method', type=str, default="NCCL")
parser.add_argument('--worker-hosts', type=str, default="localhost:23456")
parser.add_argument('--worker-index', type=int, default=0)

args = parser.parse_args()

worker_num = configure_cluster(args.worker_hosts, args.worker_index)
batch_size = args.batch_size_per_replica * worker_num
print('Batch Size: %d' % batch_size)

gpus = tf.config.experimental.list_physical_devices('GPU')
print("Physical GPU Devices Num:", len(gpus))
for gpu in gpus:
    tf.config.experimental.set_memory_growth(gpu, True)

if args.worker_method == "AUTO":
    communication = tf.distribute.experimental.CollectiveCommunication.AUTO
elif args.worker_method == "RING":
    communication = tf.distribute.experimental.CollectiveCommunication.RING
else:
    communication = tf.distribute.experimental.CollectiveCommunication.NCCL

strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy(
    communication=communication)


# logical_gpus = tf.config.experimental.list_logical_devices('GPU')
# print("Logical GPU Devices Num:", len(gpus))


def resize(image, label):
    image = tf.image.resize(image, [128, 128]) / 255.0
    return image, label


# if as_supervised is True，return image abd label
dataset, info = tfds.load("tf_flowers", split=tfds.Split.TRAIN, with_info=True, as_supervised=True)
dataset = dataset.map(resize).repeat().shuffle(1024).batch(batch_size)

# options = tf.data.Options()
# options.experimental_distribute.auto_shard = False
# dataset = dataset.with_options(options)

def build_and_compile_cnn_model():
    model = tf.keras.Sequential([
        tf.keras.layers.Conv2D(32, [3, 3], activation='relu'),
        tf.keras.layers.Conv2D(64, [3, 3], activation='relu'),
        tf.keras.layers.MaxPooling2D(pool_size=(2, 2)),
        tf.keras.layers.Dropout(0.25),
        tf.keras.layers.Flatten(),
        tf.keras.layers.Dense(128, activation='relu'),
        tf.keras.layers.Dropout(0.5),
        tf.keras.layers.Dense(info.features['label'].num_classes, activation='softmax')
    ])
    model.compile(
        opt=tf.keras.optimizers.Adam(learning_rate=0.0001),
        loss=tf.keras.losses.sparse_categorical_crossentropy,
        metrics=[tf.keras.metrics.sparse_categorical_accuracy]
    )
    return model


with strategy.scope():
    multi_worker_model = build_and_compile_cnn_model()
print("Now training the distributed model")


class TimeHistory(tf.keras.callbacks.Callback):
    def on_train_begin(self, logs={}):
        self.times = []
        self.totaltime = time.time()

    def on_train_end(self, logs={}):
        self.totaltime = time.time() - self.totaltime

    def on_epoch_begin(self, batch, logs={}):
        self.epoch_time_start = time.time()

    def on_epoch_end(self, batch, logs={}):
        self.times.append(time.time() - self.epoch_time_start)


time_callback = TimeHistory()
steps_per_epoch = 100
print('Running benchmark...')
multi_worker_model.fit(dataset, steps_per_epoch=steps_per_epoch, epochs=args.num_epochs, callbacks=[time_callback])
per_epoch_time = np.mean(time_callback.times[1:])
print("per_epoch_time:", per_epoch_time)
img_sec = batch_size * steps_per_epoch / per_epoch_time
print("Result:  {:.1f} pic/sec".format(img_sec))

TEST 2 : 작업자 1 명만, 440pic / sec （batch_szie = 128）

테스트 1 : 작업자 2 명, 610 pic / sec （batch_szie = 128 * 2） [예상 440 * 2 = 800+]

질문 1:
dist MultiWorkerMirroredStrategy 작업자 수> 1, 교육이 왜 그렇게 느린 지

배고 있다