阅读背景:

aws ec2 tensorflow gpu无法正常工作

来源:互联网 

I have one aws EC2 (p2.xlarge) with AMI

我有一个带有AMI的aws EC2(p2.xlarge)

Deep Learning AMI (Ubuntu) Version 5.0 - ami-7336d50e

深度学习AMI(Ubuntu)5.0版 - ami-7336d50e

Comes with latest binaries of deep learning frameworks pre-installed in separate virtual environments: MXNet, TensorFlow, Caffe, Caffe2, PyTorch, Keras, Chainer, Theano and CNTK. Fully-configured with NVidia CUDA, cuDNN and NCCL

附带预装在不同虚拟环境中的最新深度学习框架二进制文件:MXNet,TensorFlow,Caffe,Caffe2,PyTorch,Keras,Chainer,Theano和CNTK。完全配置NVidia CUDA,cuDNN和NCCL

I try to make rnn with keras put when a start my program i have this

当我启动我的程序时,我尝试使用keras制作rnn

 I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcublas.so.7.5 locally
 I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcudnn.so.5 locally
 I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcufft.so.7.5 locally
 I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcuda.so.1 locally
 I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcurand.so.7.5 locally

after when karas start i have this

在karas开始之后,我有了这个

W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:910] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
I tensorflow/core/common_runtime/gpu/gpu_device.cc:885] Found device 0 with properties:
name: Tesla K80
major: 3 minor: 7 memoryClockRate (GHz) 0.8235
pciBusID 0000:00:1e.0
Total memory: 11.17GiB
Free memory: 11.10GiB
I tensorflow/core/common_runtime/gpu/gpu_device.cc:906] DMA: 0
I tensorflow/core/common_runtime/gpu/gpu_device.cc:916] 0:   Y
I tensorflow/core/common_runtime/gpu/gpu_device.cc:975] Creating TensorFlow device (/gpu:0) -> (device: 0, name: Tesla K80, pci bus id: 0000:00:1e.0)
I tensorflow/core/common_runtime/gpu/pool_allocator.cc:247] PoolAllocator: After 12639 get requests, put_count=6277 evicted_count=1000 eviction_rate=0.159312 and unsatisfied allocation rate=0.590395
I tensorflow/core/common_runtime/gpu/pool_allocator.cc:259] Raising pool_size_limit_ from 100 to 110

but when de program learn is not fast my macbookpro is faster than my EC2 and i have this warning after each epochs

但是当程序学习不快时,我的macbookpro比我的EC2快,我在每个时代之后都有这个警告

tensorflow/core/common_runtime/gpu/pool_allocator.cc:247] PoolAllocator: After 4156 get requests, put_count=8233 evicted_count=4000 eviction_rate=0.48585 and unsatisfied allocation rate=0.000481232

i have install karas_gpu and tensorflow_gpu and i use vm for keras2 with tensorflow

我已经安装了karas_gpu和tensorflow_gpu,我使用vm for keras2和tensorflow

what you can tell me if I do something wrong so that a simple little macbook can be more faster than EC2 with this spec

你可以告诉我,如果我做错了什么,这样一个简单的小macbook可以比EC2更快地使用这个规范

p2.xlarge (11.75 ECU, 4 vCPU, 2.7 GHz, E5-2686v4, 61 Gio mémoire, EBS uniquement)

p2.xlarge(11.75 ECU,4个vCPU,2.7 GHz,E5-2686v4,61Giomémoire,EBS独特)

1 个解决方案

#1


0  

The response is simple. In the EC2 AMI (p2.xlarge) the gpu is Tesla K80 in with TensorFlow this gpu speedups 4x ~ 10x cpu and in my macbook i have 8 cpu.

反应很简单。在EC2 AMI(p2.xlarge)中,gpu是Tesla K80 in TensorFlow,这个gpu加速4x~10x cpu,在我的macbook中我有8个cpu。


分享到: