Friday 13 September 2019

Installing Tensorflow GPU on Fedora Linux

Following on from my previous notes on building Tensorflow for a GPU on Fedora, I find myself back at it again.  I recently upgraded my GPU at home and time has moved on too so this is my current set of notes for what I'm doing with Tensorflow on Fedora.  This method, however, differs from my previous notes in as much as I'm using the pre-built Tensorflow rather than building my own.  I've found that Tensorflow is so brittle during the build process it's much easier to work with pre-built binaries and set up my system to match their build.

In my previous blog post I benchmarked the CPU versus GPU using the Keras MNIST CNN example and so I thought it would be interesting to offer the same for this new install on my home machine.  The results are  :
  • 12 minutes and 14 seconds on my CPU
  • 1 minutes and 14 seconds on my GPU
That's just over 9.9 as fast on my GPU as my CPU!

Some info on my machine and config:
  • Custom Built Home PC
  • Intel Core i5-3570K CPU @ 3.40GHz (4 cores)
  • 16GB RAM
  • NVidia GeForce GTX 1660 (CUDA Compute Capability 7.5)
  • Fedora 30 Workstation running kernel 5.2.9-200.fc30.x86_64
Background Information for NVidia Drivers
Previously, I've always used the Negativo17 repository for all my NVidia driver and CUDA needs.  However, the software versions available there are too up-to-date to allow Tensorflow GPU to be installed in a way that works.  This repository provides CUDA 10.1 where as Tensorflow, currently at version 1.14, only supports CUDA 10.0.  So we must use another source for the NVidia software that provides back-level versions.  Fortunately, there is an official NVidia repository providing drivers and CUDA for Linux, so let's use that since it also works quite nicely with the RPM Fusion repositories as well.  Hence, this method relies purely on RPM Fusion and the official NVidia repository and does not require or use the Negativo17 repository (although it would be possible to do so).

Install Required NVidia Driver
The RPM Fusion NVidia instructions can be used here for more detail, but in brief simply install the display drivers:
  • dnf install xorg-x11-drv-nvidia akmod-nvidia xorg-x11-drv-nvidia-cuda
There are some other bits you might want from this repository as well such as:
    • dnf install vdpauinfo libva-vdpau-driver libva-utils nvidia-modprobe
    Wait for the driver to build and reboot to get things up and running.

    Install Required NVidia CUDA and Machine Learning Libraries
    This step relies on using the official nvidia repositories with a little more information available in the RPM Fusion CUDA instructions.

    First of all, add a new yum configuration file.  Copy the following to /etc/yum.repos.d/nvidia.repo:

    [nvidia-cuda]
    name=nvidia-cuda
    enabled=1
    gpgcheck=1
    gpgkey=http://developer.download.nvidia.com/compute/cuda/repos/fedora27/x86_64/7fa2af80.pub
    exclude=akmod-nvidia*,kmod-nvidia*,*nvidia*,nvidia-*,cuda-nvidia-kmod-common,dkms-nvidia,nvidia-libXNVCtrl

    [nvidia-machine-learning]

    name=nvidia-machine-learning
    baseurl=http://developer.download.nvidia.com/compute/machine-learning/repos/rhel7/x86_64/
    enabled=1
    gpgcheck=1
    gpgkey=http://developer.download.nvidia.com/compute/machine-learning/repos/rhel7/x86_64/7fa2af80.pub
    exclude=libcudnn7*.cuda10.1,libnccl*.cuda10.1



    Note that the configuration above deliberately targets the fedora27 repository from NVidia.  This is because it is the location at which we can find CUDA 10.0 compatible libraries rather than CUDA 10.1 libraries that will be found in later repositories.  So the configuration above is likely to need to change over time but essentially the message here is that we can match the version of CUDA required by targeting the appropriate repository from NVidia.  These libraries will be binary compatible with future versions of Fedora so this action should be safe to do for some time yet.



    With the following configuration in place we can now install CUDA 10.0 and the machine learning libraries required for Tensorflow GPU support and all of the libraries get installed in the correct places that Tensorflow expects.

    To install, run:
    • dnf install cuda libcudnn7 libnccl

    Install Tensorflow GPU
    The final piece of the puzzle is to install Tensorflow GPU which is now as easy as:
    • pip3 install tensorflow-gpu

    Friday 23 August 2019

    Migrating to Gnome 3

    I'm a massive laggard in the move to a Gnome 3 desktop.  Colleagues and friends have been using it for years and to be honest, I've never been comfortable using it.  But, that changed recently and I've actually grown to quite like the new desktop environment I find myself working in on a daily basis.  So I've made a full-blooded leap to a modern desktop.

    Way-back when I started using Linux as a serious desktop alternative to Windows (in about 2000-2001 ish) I was running Gnome.  I migrated away from that to KDE 3 and switched to Gnome 2 when KDE 4 was released as I didn't like the changes they had made and the new KDE 4 desktop was horribly buggy and unstable in my experience.  (Maybe there's something about brand new desktops and my not taking a liking to them?)  When Gnome released Gnome 3 I absolutely hated the user experience and used XFCE for a while before settling on the MATE desktop which I've been using for quite a few years now.

    Trying out Gnome 3 again recently and I was pleasantly surprised that the desktop has progressed significantly since those first few releases I couldn't get along with.  But it's the addition of extensions that are the final straw in my move as I've found with just the right mix I can craft a desktop that gives me a nice balance between the new world and the old, much more familiar, world.

    So, the real purpose of this post is to share the extensions I've discovered.  I'll document these below in brief but would also be interested to find others that are useful:

    Applications Menu (updated 5th Feb 2021 - no longer in use, see below)
    This was right at the very top of my list of requirements for Gnome 3 usability.  It simply puts an old school applications menu in the top bar, a bit like your old fashioned Windows start menu or similar from other desktops.  I am, however, finding I use this very little now as the search hot-key in Gnome 3 does seem to be a quicker way of finding and starting programs.

    Frippery Bottom Panel
    This is another of my top requirements for Gnome 3 usability.  It gives you a panel at the bottom of the screen (D'uh) that allows you to switch easily between applications you have running.  It also has a small workspace switcher which is why I like the Frippery version of this type of extension versus some of the others that don't have a workspace switcher capability.

    Top Icons Plus (update 5th Feb 2021 - no longer in use, replaced with AppIndicator, see below)
    Either the Top Icons or the Top Icons Plus extension that I'm using here seem so ubiquitous for Gnome 3 users I wonder why on earth they're not a default option, aside from the fact the Gnome 3 developers do seem to retain their keen vision on what a modern desktop should look like and "old" system tray icons are not part of that outlook.  This extension, if you're not already using it, allows you to see system tray icons such as the ones used by Virt Manager or Slack, for example.

    GPaste (update 4th Nov 2021 - no longer in use, too difficult to configure and doesn't always work the way I expected, replaced with Clipboard Indicator, updated again 8th Mar 2023, replaced with Pano Clipboard Manager)
    A clipboard management system that has a nice integration with the Gnome 3 panel.  I was previously using apps like ClipIt or Parcelite that do pretty much the same job.

    Lock Screen (update 5th Feb 2021 - no longer in use, I wasn't using this as I just hit Win+L to lock)
    This adds a button to the gnome panel that, when clicked, locks your desktop.  This would be the same as pressing Win+L on the keyboard.  I was in the habit of using a graphical button on MATE so having this back in Gnome 3 gives me the experience I'm used to.

    No TopLeft Hot Corner (update 5th Feb 2021 - no longer in use, Gnome Tweaks as a toggle for "Activities Overview Hot Corner" from the Top Bar options.
    I find the Gnome 3 facility to show activities when you mouse to the top left corner really annoying and it detracts from my productivity when it happens automatically.  Fortunately, this extension disables that feature.  It does make it more awkward to reach activities with the mouse (I'd have to click the applications menu first then select "Activities Overview") but I more or less always use the Windows key anyway.

    Places Status Indicator (update 8th Mar 2023 - no longer in use, I wasn't using it so have stopped installing it)
    This adds the old Gnome 2 style places menu to the Gnome 3 panel.  I find I flip between using this menu to start navigating directories and just starting Gnome Files and going from there.  Any which way, having this menu back on my desktop just makes it feel a bit more familiar and comfortable.

    Remove Dropdown Arrows (update 5th Feb 2021 - no longer in use, Gnome seems to have gotten rid of most of these by default)
    The Gnome 3 panel insists on having an arrow indicator to show items that pull down a menu when clicked.  These menus seem obvious to me and the arrows look rubbish and take up space, so this extension gets rid of them completely.  Happy days.

    Suspend Button (update 5th Feb 2021 - no longer in use, Gnome now has a built-in suspend button)
    I run from a laptop most of the time and use the suspend feature every time I "shut down" my laptop.  Bizarrely, there's no graphical facility (that I can find) in Gnome to suspend my machine.  This extension adds a nice button to the status menu that immediately suspends my machine.  Perfect.

    System Monitor
    Adds little graphs to the Gnome panel that show resource usage.  The extension is pretty configurable but I have it showing CPU, memory and network utilisation.  This allows me to keep an easy eye on my machine and how loaded it is at the current time.  Extremely useful for spotting those occasional rogue apps that start eating an entire core of my CPU.

    Media Keys (update 5th Feb 2021 - no longer in use, Gnome has media controls built-in)
    I haven't decided how useful this one is going to be yet and it's currently turned off.  However, when listening to Music through services like Amazon Music from a web browser it's nice to be able to control the audio without having to revert back to the browser ever time.  This extension simply adds a few buttons to the Gnome panel to control your media.  Handy if you haven't got the physical buttons on your keyboard too.

    Do Not Disturb Button (update 5th Feb 2021 - no longer in use, Gnome has the button built-in)
    I generally leave this extension disabled but it's useful to have installed and running when presenting or screen sharing.  It saves any embarrassing situations of people being able to read your notifications while they're looking at your screen.  Basically, it simply stops notifications being displayed, they're still received so you can go read them later.


    Blog edited with more extensions added on 28th August 2019:
    Frippery Panel Favourites
    I'm not quite sure how I missed this from my original list as it's an extension I've been using more or less since day one in Gnome 3.  It takes your favourite menu and adds this as a set of icons to the top of the Gnome Panel.  Makes for extra quick access to your commonly used apps.

    Some more extensions have been brought to my attention since writing the list above.  I've tried out all of the ones mentioned to me but these additions (below) are the ones that seem to have stuck.

    Caffeine (update 5th Feb 2021 - no longer in use, using Gnome's built-in do not disturb button instead)
    This extension sits fairly well alongside the Do Not Disturb Button extension in my original list.  This one simply disables the screen saver and auto suspend.  Hence, in conjunction with Do Not Disturb, will make a good presentation or screen sharing environment.

    GTile
    This is a genius little extension that allows you to easily resize your windows in order to tile them across your display.  I love the side-snapping in Gnome 3 that allows you to size a window to half the screen size.  In my older desktops I also had corner snapping to size a window to a quarter of the screen, Gnome 3 doesn't have this by default.  However, GTile adds an icon to your Gnome Panel that, when clicked, allows you to size to any area of your screen across a pre-defined grid - you can even change the grid size.  Brilliant for usability with lots of on-screen windows at the same time.


    Blog edited to update the list of extensions I'm using on 5th February 2021:
    Applications Menu
    Just a note to say that after using Gnome 3 for quite some time now, I rarely (if ever) use the Applications Menu any longer.  I tend to start applications either by pressing the Gnome hot key (Windows Key by default) and type in the search box, or by clicking on one of the favourites in the panel via the "Panel Favourites" extension.

    While the free desktop standard continues to specify the classic "icon tray" that was supported by extensions such as TopIcons, in reality few of the modern desktops (I'm referring to Gnome and KDE) support them.  The modern take on the tray icon is an AppIndicator icon and many modern applications are written to use this standard (and perhaps fall back to a tray icon).

    Blog edited to update the list of extensions I'm using on 8th March 2023:
    It's interesting to see how much use of extensions has changed over time.  I'm using considerably fewer now than I was when I first started using Gnome 3.  I put this down to two things: (1) Gnome is better at operating the way user's expect by removing the need for extensions such as the Suspend Button in the list above; and (2) I've become more institutionalised to the way that Gnome works, I'm much more familiar with it and have grown to like much of the way it works.

    Wednesday 3 April 2019

    Helping Disabled Passengers Travel with Confidence

    This is a reproduction of a blog post originally made on the IBM Emerging Technology blog at https://www.ibm.com/blogs/emerging-technology/helping-disabled-passengers-travel-with-confidence/ (link now dead).  Original article published on 3rd April 2019, re-publication to this blog was on 11th June 2021.

    Introduction

    Disabled or disadvantaged passengers have every right to fear travelling with multiple high profile cases of unfair and appalling levels of service evident in the media, particularly for rail passengers. Our team have been tackling this problem recently in collaboration with the IBM Travel and Transport team and 50% funding through Innovate UK’s call for Accelerating Innovation in Rail (round 4). The project was the brain child of Sam Hopkins and won the internal IBM Hybrid Hackathon event in 2016.

    The Problem

     The rail industry in the UK is a complex one and travelling on our railways can be a difficult experience for anybody. If you’re disabled then the experience can be difficult at best and demeaning or impossible at worst. Depending on the type and severity of their disability, disabled travellers need to plan for eventualities such as whether a station is fully accessible, whether and when staff are available to help, when and where they might be able to use a toilet, how and where to board and alight the train, the list is as long as the wide range of issues people may have. Today, assistance is on offer, but this is complex due to the nature of our fragmented system and passengers (or their carers) may need to contact multiple companies 24 hours in advance of travel when planning their journey. This process of organising their journey is extremely time consuming and can be as much of a frustration as the journey itself.

    The Challenges 

    The UK rail industry is highly fragmented with different parts of the operation split among different companies. For example, Network Rail oversee the infrastructure such as maintaining tracks and signalling; Rolling Stock Operating Companies (ROSCOs) own the trains; and Train Operating Companies (TOCs) operate the trains and run passenger services in different regions around the UK. In order for a disabled passenger to make a journey, the Train Operating Companies are legally obliged to provide support both off the train at the station and on board the train during the journey. However, a passenger may be travelling through a large number of regions and across services provides and hence need to understand which company will be responsible for their journey at any given stage. Understanding this takes quite a bit of research and can be difficult, it also requires that everything they require on their journey lines up across the network they’re travelling on and this is certainly not guaranteed on our network today.

    Our Solution

    We have created a mobile application for disabled passengers, supported by an enterprise grade back end architecture. During the project we established a principle that the disabled passenger should always be connected to a member of staff. This connection provides continuous support available whenever required by the passenger and is designed to help

    alleviate the anxiety felt by this group of passengers when travelling. The connection comes in the form of a chat application similar to how passengers may already use common applications such as WhatsApp or Facebook Messenger. The passenger can type anything they want to the member of staff they’re connected to and get a human-level response to their query from staff local to their situation. During their journey, the passenger is handed over between staff members such that they’re always connected with a local staff member who can offer physical assistance if required.

    A typical journey story starts with the user programming their journey in a similar way to how they might on existing rail applications. They search for their departure and destination stations and select a time and route they wish to travel. Some time later, they embark upon their journey. When they arrive at the station, they either press a button to start the process or the app uses geolocation to detect their arrival at the station. They are connected to a member of the station staff who can answer questions and provide physical assistance to board their first train. Once boarded, the station staff hand the passenger over to a member of staff on board the train. This hand-over process is transparent to the user and they will simply see a new greeting from the next member of staff in the chain once they have been connected. The process continues with the on train staff handing over to station staff at the destination station in order to provide assistance alighting the train. This process can repeat as many times as necessary until the final destination is reached.

    Technical Implementation 

    Our solution has a strong server-side implementation with an MVP level front end to exercise the APIs provided.

    We base the solution entirely within the IBM Cloud, using a Cloudant instance for database hosting, App ID for authentication services, Message Sight for highly reliable messaging built on top of the MQTT protocol, and NodeJS for writing our APIs on top of the Express framework and where necessary calling out to other APIs such as the Transport API. Security and privacy have been key concerns in the design to ensure chat messages are securely delivered only to their intended recipient. The application itself is authenticated at each of its routes and all of the server-side APIs also use the same App ID authentication. Novel to this solution is the capability of third-party authentication from Message Sight to App ID, a first-of-a-kind implementation.

    The front end application is currently written in VueJS. This has been designed with accessibility in mind given the intended audience. It is a simple interface conforming to web accessibility standards that is compatible with screen readers. This component was never intended to be the final solution and we see another interface being developed in the future that is likely to build upon the current one. For example, a fully native solution or perhaps one written to be more native-like using technologies such as Native Script.

    Moving Forward 

    The solution is complete and ready, subject to customisation, to be adopted by the UK train operating companies. However, we do see the need for an improved user interface to be developed before the solution is truly ready to be used in the wild.

    Further to the current solution for the rail industry, we recognise the issues faced by disabled travellers don’t stop with the UK rail system. Similar issues are faced when using our road network, taxi services, ferries and of course our airports. The principle established within this project of alleviating anxiety through the connection to a supporting human member of staff is likely to remain. We hope this is extended into the other areas we’re considering with the next most obvious choice to tackle being the airport use case.

    Friday 18 January 2019

    Self-Signing SSL/TLS Certificates

    Things have changed a bit since I last looked into setting up a Certificate Authority (CA) and using that to self-sign my own certificates, not least that the use of the Common Name (CN) field appears to have changed. Chrome in particular seems to insist on the use of the Subject Alternative Names (SAN) extension rather than (or in addition to) using the CN field. So these are my notes on how to set up your own CA and use that to sign certificates. I'm conscious this is bound to go out of date so at the time of writing I'm working with Firefox 64, Chrome 71 and OpenSSL 1.1.1.

    Setup
    First of all, create a config file along the lines of the following and call it anything you like but for these notes I'm going to call it ssl.conf.  Note, if you want to you can start with a different template or look at your own openssl.cnf file which on Linux is commonly found at /etc/pki/tls/openssl.cnf.

    [ req ]
    default_bits       = 4096
    distinguished_name = req_distinguished_name
    req_extensions     = req_ext
    
    [ req_distinguished_name ]
    countryName                 = Country Name (2 letter code)
    countryName_default         = GB
    stateOrProvinceName         = State or Province Name (full name)
    stateOrProvinceName_default = England
    localityName                = Locality Name (eg, city)
    localityName_default        = MyCity
    organizationName            = Organization Name (eg, company)
    organizationName_default    = MyOrg
    commonName                  = Common Name (e.g. server FQDN or YOUR name)
    commonName_max              = 64
    commonName_default          = localhost
    
    [ req_ext ]
    subjectAltName = @alt_names
    
    [alt_names]
    DNS.1 = localhost
    

    You can change any of this template and indeed you'll need to change the common name for the certificates you're generating. The CN can be changed either on the command line during certificate creation or by changing the default in ssl.conf. You will also need to change the list of names under the "alt_names" section, this list should contain one line for each host name your machine might be known as. The list starts at DNS.1 for the first entry, then you can add DNS.2 for the second entry and so on.

    NOTE: the specification and a lot of the documents available in this space indicate that an IP address can be used in the CN.  My testing seems to indicate that while this is the case, certificates produced in this way will be rejected by modern browsers.  Hence, you should list only hostnames as the CN but IP addresses still appear to be acceptable in as "alt_names".

    Create a Certificate Authority
    You'll need a certificate and key file to act as your own CA:

    openssl genrsa -out RootCA.key 4096
    openssl req -x509 -new -nodes -key RootCA.key -sha256 -days 3650 -out RootCA.pem -config ssl.conf

    You can inspect the certificate with:
    openssl x509 -in RootCA.pem -text -noout

    Create a Certificate Signing Request (CSR)
    Now you have a CA you can create a CSR that can be used with your CA certificate to generate a client certificate:

    openssl genrsa -out server.key 4096
    openssl req -new -key server.key -out server.csr -config ssl.conf

    You can inspect the certificate with:
    openssl req -text -noout -verify -in server.csr

    This time it's really important to ensure your host names are listed under the "X509v3 Subject Alternative Name" section of the certificate.

    Generate a Signed Certificate
    You can now use the CSR to create a signed certificate that can be used to serve up content over a secure connection:

    openssl x509 -req -in server.csr -CA RootCA.pem -CAkey RootCA.key -CAcreateserial -out server.pem -days 3650 -sha256 -extensions req_ext -extfile ssl.conf

    Note: if you want to create a different format of certificate here you can simply replace server.pem in the above command with something like server.crt, for example.

    You can inspect the certificate with:
    openssl x509 -in server.pem -text -noout

    Again, it's really important to ensure your host names are listed under the "X509v3 Subject Alternative Name" section of the certificate.

    Use the Certificate Server Side
    You can now put your server.pem and server.keyfiles to work and serve up content over a secure connection.  There's too many ways to do that to list here but it could be used with a web server to serve HTTPS or a websockets server to serve some sort of socket connection as a couple of examples.

    If you want to inspect the certificate that's being used on the server, replace <hostname> and <port> in the command below:

    openssl s_client -connect <hostname>:<port> | openssl x509 -noout -text

    Use the Certificate Client Side
    My use case here is with a web browser and so you'll want to import your <i>RootCA.pem</i> into your browser environment.  There are two main ways of achieving this, you can either:
    1. Import directly to the browser
    2. Import to the key store on your operating system
    It's quicker and easier to import directly to the browser but this will of course only cover that one browser on your system whereas if you use the operating system method then any application that consults the OS for certificates will see your CA certificate.

    For Firefox, go to "View Certificates" in the preferences; click the "Authorities" tab and then the "Import" button; select your <i>RootCA.pem</i> file and click OK.

    For Chrome, go to "Manage Certificates" in the settings; click the "Authorities" tab and then the "Import" button; select your <i>RootCA.pem</i> file; click the check boxes to trust the certificate and click OK.