Jupyter Notebook上でGPUの情報を確認する方法を記載します．
目次
- 1. 実行環境
- 2. 各種確認
- 2.1. TensorFlowのライブラリからデバイス設定を確認
- 2.2. GPUの種類確認
- 2.3. Cudaのバージョン確認
- 2.4. GPUの使用状況確認
- 2.5. NVIDIA DRIVERのバージョン確認
- 3. 補足
- 4. 参考リンク
1. 実行環境
- Amazon SageMaker
- ml.p2.xlargeを利用
2. 各種確認
2.1. TensorFlowのライブラリからデバイス設定を確認
from tensorflow.python.client import device_lib
device_lib.list_local_devices()
[name: "/device:CPU:0"
device_type: "CPU"
memory_limit: 268435456
locality {
}
incarnation: 8738958628955863132, name: "/device:GPU:0"
device_type: "GPU"
memory_limit: 11285974221
locality {
bus_id: 1
links {
}
}
incarnation: 17652009205189153751
physical_device_desc: "device: 0, name: Tesla K80, pci bus id: 0000:00:1e.0, compute capability: 3.7"]
CPUが利用できることを確認できた．
2.2. GPUの種類確認
%%bash
lspci | grep -i nvidia
00:1e.0 3D controller: NVIDIA Corporation GK210GL [Tesla K80] (rev a1)
Tesla K80が確認出来る
2.3. Cudaのバージョン確認
%%bash
nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Sep__1_21:08:03_CDT_2017
Cuda compilation tools, release 9.0, V9.0.176
Cudaのバージョンは9.0
2.4. GPUの使用状況確認
%%bash
nvidia-smi
Mon Jul 9 23:09:59 2018
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 384.111 Driver Version: 384.111 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla K80 On | 00000000:00:1E.0 Off | 0 |
| N/A 44C P0 70W / 149W | 10875MiB / 11439MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 7302 C ...naconda3/envs/tensorflow_p27/bin/python 10862MiB |
+-----------------------------------------------------------------------------+
2.5. NVIDIA DRIVERのバージョン確認
%%bash
cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module 384.111 Tue Dec 19 23:51:45 PST 2017
GCC version: gcc version 7.2.1 20170915 (Red Hat 7.2.1-2) (GCC)
3. 補足
Jupyter Notebookでlinux commandは下記の通り実行できます．
当たり前ですが，どちらも出力結果は同じ．
実行方法1
%%bash
nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Sep__1_21:08:03_CDT_2017
Cuda compilation tools, release 9.0, V9.0.176
実行方法2
!nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Sep__1_21:08:03_CDT_2017
Cuda compilation tools, release 9.0, V9.0.176