GrooveYOU

 참고 사이트

http://hwengineer.blogspot.com

http://shareithw.blogspot.kr

http://leearmykr.blogspot.com/

1. Download the ISO file for the Linux distribution and create a bootable USB device or DVD.
2. Insert your USB device into the front USB port.
3. Verify that USB device is shown as an option in Petitboot as USB: sdc:

   USB: sdc /
   		Rescue mode
   		Install

   Note:Select Rescan devices if USB device does not appear.

   If you are installing Red Hat Enterprise Linux 7.2, you must provide some additional details to the installer. Follow these steps:

   1. Record the UUID of the USB device. For example, the UUID of the USB device in the following example is 2015-10-30-11-05-03-00.

      [USB: sdb1 / 2015-10-30-11-05-03-00]
                  Rescue a Red Hat Enterprise Linux system (64-bit kernel)
                  Test this media & install Red Hat Enterprise Linux 7.2  (64-bit kernel)
               *  Install Red Hat Enterprise Linux 7.2 (64-bit kernel)

   2. Select Install Red Hat Enterprise Linux 7.2 (64-bit kernel) and press e (Edit) to open the Petitboot Option Editor window.
   3. Move the cursor to the Boot arguments section and add the following information:

      inst.stage2=hd:UUID=your_UUID

      where your_UUID is the UUID that you recorded.

      Petitboot Option Editor
                    qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
      
                    Device:    ( ) sda2 [f8437496-78b8-4b11-9847-bb2d8b9f7cbd]
                                    (*) sdb1 [2015-10-30-11-05-03-00]
                                    ( ) Specify paths/URLs manually
      
                    Kernel:         /ppc/ppc64/vmlinuz
                    Initrd:         /ppc/ppc64/initrd.img
                    Device tree:
                    Boot arguments: ro inst.stage2=hd:UUID=2015-10-30-11-05-03-00    
      
                       [    OK    ]  [   Help   ]  [  Cancel  ]

   4. Select OK to save your options and return to the Main menu.
4. Select Install.
   Note:Be patient! It can sometimes take a couple minutes for the installation to begin.
5. Follow the installation wizard for your Linux distribution to set up disk options, your user name and password, time zones, and so on. The last step is to restart your system.
6. After the system restarts, Petitboot displays the option to boot the Linux distribution that you installed. Select this option and press Enter.

### Operating System and Repository Setup

   1. Enable 'optional' and 'extra' repo channels

          $ sudo subscription-manager repos --enable=rhel-7-for-power-9-optional-rpms

          $ sudo subscription-manager repos --enable=rhel-7-for-power-9-extras-rpms

   2. Install packages needed for the installation

          $ sudo yum -y install wget nano bzip2

   3. Enable EPEL repo ( 위에 설치함 생략 )

          $ wget https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm

          $ sudo rpm -ihv epel-release-latest-7.noarch.rpm

   4. Load the latest kernel

          $ sudo yum update kernel kernel-tools kernel-tools-libs kernel-bootwrapper

          $ reboot # This reboot may be deferred until after the NVIDIA steps below.

      Or do a full update

          $ sudo yum update

          $ sudo reboot # This reboot may be deferred until after the NVIDIA steps below.

### NVIDIA Components

Before installing the NVIDIA components the udev Memory Auto-Onlining Rule must

be disabled for the CUDA driver to function properly. To disable it:

   1. Edit the /lib/udev/rules.d/40-redhat.rules file.

          $ sudo nano /lib/udev/rules.d/40-redhat.rules

   2. Comment out the following line and save the change:

          SUBSYSTEM=="memory", ACTION=="add", PROGRAM="/bin/uname -p", RESULT!="s390*", ATTR{state}=="offline", ATTR{state}="online"

   3. Reboot the system for the changes to take effect.

          $ sudo reboot

Download IBM Spectrum MPI from the ESP download site.

   1. Install the rpms

          $ sudo rpm -ihv ibm_smpi_lic_s-10.02.00*.ppc64le.rpm ibm_smpi-10.02.00*.ppc64le.rpm

### Software Repository Setup

IBM TensorFlow ESP for Power AC922 Deep Learning packages are distributed in an rpm file and is available from the ESP download site.

Installing the rpm creates an installation repository on the local machine.

   1. Install the repository package:

          $ sudo rpm -ihv mldl-repo-local*.rpm

### Installing all frameworks at once

All the Deep Learning frameworks can be installed at once using the

`power-mldl` meta-package:

        $ sudo yum install power-mldl-esp

### Installing frameworks individually  

 -> 위에 power-mldl-esp 로 설치 했으면 SKIP!!!!!!

The Deep Learning frameworks can be installed individually if preferred.

The framework packages are:

   - `tensorflow` - Google TensorFlow, v1.4.0

   - `tensorboard` - Web Applications for inspecting TensorFlow runs and graphs, v0.4.0rc3

   - `ddl-tensorflow` - Distributed Deep Learning custom operator for TensorFlow

Each can be installed with:

        $ sudo yum install <framework>-cuda9.1

### Accept the License Agreement

Read the license agreements and accept the terms and conditions before using Spectrum MPI or any of the frameworks.

        $ sudo IBM_SPECTRUM_MPI_LICENSE_ACCEPT=no /opt/ibm/spectrum_mpi/lap_se/bin/accept_spectrum_mpi_license.sh

        $ sudo /opt/DL/license/bin/accept-powerai-license.sh

After reading the license agreements, future installs may be automated to silently accept the license agreements.

        $ sudo IBM_SPECTRUM_MPI_LICENSE_ACCEPT=yes /opt/ibm/spectrum_mpi/lap_se/bin/accept_spectrum_mpi_license.sh

        $ sudo IBM_POWERAI_LICENSE_ACCEPT=yes /opt/DL/license/bin/accept-powerai-license.sh

## Tuning Recommendations

Recommended settings for optimal Deep Learning performance on the IBM Power System AC922 are:

   - Enable Performance Governor

         $ sudo yum install kernel-tools

         $ sudo cpupower -c all frequency-set -g performance

   - Enable GPU persistence mode

         $ sudo systemctl enable nvidia-persistenced

         $ sudo systemctl start nvidia-persistenced

   - For TensorFlow, set the SMT mode

         $ sudo ppc64_cpu --smt=4

   - For TensorFlow with DDL, set the SMT mode

         $ sudo ppc64_cpu --smt=2

## Getting Started with MLDL Frameworks

### General Setup

Most of the PowerAI packages install outside the normal system search

paths (to `/opt/DL/...`), so each framework package provides a shell

script to simplify environmental setup (e.g. `PATH`, `LD_LIBRARY_PATH`,

`PYTHONPATH`).

We recommend users update their shell rc file (e.g. `.bashrc`) to source

the desired setup scripts. For example:

    $ source /opt/DL/<framework>/bin/<framework>-activate

Each framework also provides a test script to verify basic function:

    $ <framework>-test

### Note about dependencies

A number of the PowerAI frameworks (for example, TensorFlow, and TensorBoard)

have their dependencies satisfied via Anaconda packages.  These dependencies are validated

by the `<framework>-activate` script to ensure they are installed and, if not, the script will fail.

For these frameworks, the `/opt/DL/<framework>/bin/install_dependencies` script must be run

prior to activation to install the required packages.

For example:

    $ source /opt/DL/tensorflow/bin/tensorflow-activate

    Missing dependencies ['backports.weakref', 'mock', 'protobuf']

    Run "/opt/DL/tensorflow/bin/install_dependencies" to resolve this problem.

    $ /opt/DL/tensorflow/bin/install_dependencies

    Fetching package metadata ...........

    Solving package specifications: .

    Package plan for installation in environment /home/rhel/anaconda2:

    The following NEW packages will be INSTALLED:

        backports.weakref: 1.0rc1-py27_0

        libprotobuf:       3.4.0-hd26fab5_0

        mock:              2.0.0-py27_0

        pbr:               1.10.0-py27_0

        protobuf:          3.4.0-py27h7448ec6_0

    Proceed ([y]/n)? y

    libprotobuf-3. 100% |###############################| Time: 0:00:02   2.04 MB/s

    backports.weak 100% |###############################| Time: 0:00:00  12.83 MB/s

    protobuf-3.4.0 100% |###############################| Time: 0:00:00   2.20 MB/s

    pbr-1.10.0-py2 100% |###############################| Time: 0:00:00   3.35 MB/s

    mock-2.0.0-py2 100% |###############################| Time: 0:00:00   3.26 MB/s

    $ source /opt/DL/tensorflow/bin/tensorflow-activate

    $

The TensorFlow homepage

([https://www.tensorflow.org/](https://www.tensorflow.org/)) has a

variety of information, including Tutorials, How Tos, and a Getting

Started guide.

Additional tutorials and examples are available from the community, for

example:

   - [https://github.com/nlintz/TensorFlow-Tutorials](https://github.com/nlintz/TensorFlow-Tutorials)

   - [https://github.com/aymericdamien/TensorFlow-Examples](https://github.com/aymericdamien/TensorFlow-Examples)

#### Distributed Deep Learning (DDL) Custom Operator for TensorFlow

IBM TensorFlow ESP for Power AC922 includes a Technology Preview of the

IBM PowerAI Distributed Deep Learning (DDL) custom operator for TensorFlow.

The DDL custom operator uses IBM Spectrum MPI and NCCL to provide high-speed

communications for distributed TensorFlow.

The DDL custom operator can be found in the `ddl-tensorflow` package.

For more information about DDL and about the TensorFlow operator, see:

   - `/opt/DL/ddl-doc/doc/README.md`

   - `/opt/DL/ddl-tensorflow/doc/README.md`

   - `/opt/DL/ddl-tensorflow/doc/README-API.md`

The DDL TensorFlow operator makes it easy to enable models

for distribution. The package includes examples of models enabled

with DDL including TensorFlow High Performance and Slim models:

    $ source /opt/DL/ddl-tensorflow/bin/ddl-tensorflow-activate

    $ ddl-tensorflow-install-samples <somedir>

The Slim model examples are based on a specific commit of the TensorFlow models

repo with a small adjustment. If you prefer to work from an upstream clone,

rather than the packaged examples:

    $ git clone https://github.com/tensorflow/models.git

    $ cd models

    $ git checkout 11883ec6461afe961def44221486053a59f90a1b

    $ git revert fc7342bf047ec5fc7a707202adaf108661bd373d

    $ cp /opt/DL/ddl-tensorflow/examples/slim/train_image_classifier.py slim/

#### Additional TensorFlow Features

The PowerAI TensorFlow packages include TensorBoard. See:

[https://www.tensorflow.org/get_started/summaries_and_tensorboard](https://www.tensorflow.org/get_started/summaries_and_tensorboard)

The TensorFlow 1.4.0 package includes support for additional features:

   - HDFS

   - NCCL

   - experimental XLA JIT compilation

     (see [https://www.tensorflow.org/performance/xla/](https://www.tensorflow.org/performance/xla/))

## Uninstalling MLDL Frameworks

The MLDL Framework packages can be uninstalled individually the same way they were installed. In order to uninstall all MLDL packages and the repo used to install them run:

    $ sudo yum remove powerai-license

    $ sudo yum remove mldl-repo-local-esp

1. Redhat 환경의 경우

[root@ac922 ~]# cat /etc/yum.repos.d/docker.repo
[docker]
name=Docker
baseurl=http://ftp.unicamp.br/pub/ppc64el/rhel/7/docker-ppc64el/
enabled=1
gpgcheck=0

# systemctl enable docker.service
# systemctl start docker.service
-> Docker 서비스 시작!

# docker images

-> Docker 실행 확인

///////////////////////////////////////////////////////////////////////////////

[이하 nvidia-docker]

RHEL-based distributions

# distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
# curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.repo | sudo tee /etc/yum.repos.d/nvidia-docker.repo

-> Nvidia-docker repository 업데이트

# yum update

# yum install nvidia-docker

# nvidia-docker

-> nvidia-docker 설치 확인

# /usr/bin/nvidia-docker-plugin &

-> nvidia-docker 사용을 위해 plugin 을 background로 실행

-> 재부팅후에도 background 로 실행하기 위해 rc.local 에 등록

# vi /etc/rc.d/rc.local

/usr/bin/nvidia-docker-plugin &

추가해줌

# chmod +x /etc/rc.d/rc.local

-> rc.local에 실행권한 필수

Disabling the Subscription-Manager Repository

When a system is registered using Subscription-Manager, the rhsmcertd process creates a special yum repository — redhat.repo. As “Enabling Supplementary and Optional Repositories” describes, as the system adds subscriptions, the product channels are added to the redhat.repo file.

Maintaining a redhat.repo file may not be desirable in some environments. It can create static in content management operations if that repository is not the one actually used for subscriptions, such as for a disconnected system or a system using a local content mirror.

This default redhat.repo repository can be disabled by editing the Subscription-Manager configuration and setting the manage_repos value to zero (0).

Raw

[root@server ~]# subscription-manager config --rhsm.manage_repos=0