Planet GridPP

Software Sustainability Institute Blog | The "little book of software struggles" and other great ideas from our Fellows

2013-06-19T08:50:22+00:00

Every six months we meet with our Fellows to discuss their experiences of Fellowship programme. Last week's Fellows' meeting had a very busy schedule! Almost 24 intensive hours (with a break for a good night sleep) filled with discussions, ideas and experiences.

Software Sustainability Institute Blog | Monkeying around with chemistry: the APES project

2013-06-18T09:46:46+00:00

By Arno Proeme.

A version of this post originally appeared on the EPCC blog.

Unfortunately, the APES project (pronounced A-PES) has nothing to do with our chimp cousins, and instead stands for Advanced Potential Energy Surfaces. It will help researchers advance their understanding of the structure and function of molecules by improving the models used to represent their force fields. For example, it could help understand the interaction between the simian olfactory receptors and isoamyl acetate, which gives bananas their irresistible allure.

WLCG RAL Tier 1 | LSI-1068E issues understood and “resolved”

2013-06-14T15:16:23+00:00

TL;DR version: lsi 1068e is not suitable for workloads involving 4 drives, constantly deleting and writing thousands of small files (aprox. 320,000 in our case) in RHEL5/SL5

Back in early 2011 several Dell R410 were purchased, one of them destined to be a development box for Quattor work. While most of the R410s had 2 drives in RAID1, that one had 4 identical drives in 2 RAID1 arrays, , and then software RAID0 on top of these 2 RAID1 arrays. It also had faster CPUs and memory.

It soon turned out the machine would not behave : after a random time working fine, it would start slow down, spit various disk related errors, and eventually crash. Rebooting it would fix the issues temporary.

In the course of the following 2 years everything was replaced in that machine: disks, CPUS, motherboard, backplane, SAS cables. No change. In early 2013 the decision was made to replace the entire box with a new one, identical the old one. Still no change: new and old machine behaved exactly the same.

It became rather obvious we were dealing with a software or firmware/hardware bug.

Detailed systems specs

TBD

Errors samples

TBD

New drivers

Default RHEL/SL drivers for lsi mpt sas are rather old, so the latest drivers were downloaded and tested from LSI website. Dell drivers were slightly older than the LSI ones but the were tested too. There was no change

Additional debugging

TBD

RHEL vs SL

TBD

Bug replication

Triggering the bug was not easy as under normal system function it could occur anything between 2 weeks to few hours of uptime. Normal disk server acceptance tests would hammer the machine for 10+ days without any sign of problem.

Software Sustainability Institute Blog | Software and polar research workshop

2013-06-14T13:19:38+00:00

By Allen Pope, postdocoral student and Fellow.

With sponsorship from the Software Sustainability Institute, the UK Polar Network (UKPN) are running a Software & Polar Research Workshop at the Scott Polar Research Institute, Cambridge on Tuesday 17 September 2013. The workshop is currently in the planning process, but please save the date in your diaries and check back regularly for updates!

Where

Scott Polar Research Institute, University of Cambridge, Lensfield Road, Cambridge, CB2 1ER.

Software Sustainability Institute Blog | Software Carpentry at Southampton

2013-06-10T09:41:40+00:00

By Mike Jackson.

Last week saw the first Software Carpentry boot camp to be held at the University of Southampton. Organised by our fellow, Robin Wilson, fifty researchers from the Institute for Complex Systems Simulation and Computational Modelling Group, attended this boot camp, with Nelle Varoquax and James Morrison (who attended our Manchester boot camp back in April) as instructors.

Institute staff Steve Crouch and Tim Parkinson provided local help. Pop over to the Software Carpentry blog to read about how the boot camp went, and check out Six useful things I have learnt at Software Carpentry boot camp written by a satisfied attendee.

Software Sustainability Institute Blog | Top tips for working with a technical writer

2013-06-07T07:55:51+00:00

By Craig Haiss, HelpScribe.

If you want people to use your research software, it needs good documentation. Many people employ the help of a technical writer to write documentation, but how do you get the best out of this relationship? We turned to Craig Haiss for help. He set up HelpScribe to share his ideas with the technical writing community, and in 2010 it was named one of the most influential technical writing blogs on the web.

1. Accept that no tool or process is as simple as it seems

The more functionality a tool provides, the greater the chances a user will become confused by features of that tool. Even the iPhone, with it's simple touch screen interface, has a 150+ page manual.

Do not assume that your research software or process is so simple that minimal or no documentation is sufficient. Doing so could cost you both time and money when research grinds to a halt due to unanswered questions or misunderstandings.

GridPP storage news | Demonstrating 100 Gb/s

2013-06-06T04:06:22+00:00

One of the interesting things from a networkingdataological perspective at this year's TNC is showing 100 Gb/s link across the Atlantic, and also the conference itself is connected with 100 Gb. 100 Gb is here!

We also had a music performance with a local (Maastricht) band and one musician in Edinburgh, and they were playing together. This can of course only happen if the latency is very low - 10-20ms. First time I have seen this in practice, very impressive.

LondonGrid | Serial Consoles over ipmi

2013-06-04T15:00:03+00:00

To get Serial Consoles over ipmi working properly with Scientific Linux 6.4 (aka RHEL 6.4 / centos 6.4) I had to modify several setting both in the BIOS and in the OS.

Hardware Configuration

For Dell C6100 I set these setting in the BIOS

Remote Access = Enabled
Serial Port Number = COM2
Serial Port Mode = 115200 8,n,1
Flow Control = None
Redirection After BIOS POST = Always
Terminal Type = VT100
VT-UTF8 Combo Key Support = Enabled

Note: "Redirection After Boot = Disabled" is required otherwise I get a 5 minute timeout before booting the kernel. Unfortunately with this set up you get a gap in output while the server attempts to pxeboot. However, you can interact with the BIOS and once Grub starts you will see and be able to interact with the grub and Linux boot processes.

For Dell R510/710 I set these setting in the BIOS

Serial Communication = On with Console Redirection via COM2
Serial Port Address = Serial Device1=COM1,Serial Device2=COM2
External Serial Connector = Serial Device1
Failsafe Baud Rate = 115200
Remote Terminal Type = VT100/VT220
Redirection After Boot = Disabled

Note: With these settings you will be unable to see the progress of the kickstart install on the non default console.

Grub configuration

In grub.conf you should have these two lines (they were there by default in my installs).

serial --unit=1 --speed=115200
terminal --timeout=5 serial console

This allows you access grub via the consoles. The "serial" (ipmi) terminal will be default unless you press a key when asked during the boot process. This is only for grub and not for the rest of the linux boot process

SL6 Configuration

The last console specified in the linux kernel boot options is taken to be the default console. However, if the same console is specified twice this can cause issues (e.g. when entering a password the characters are shown on the screen!)

For the initial kickstart pxe boot I append "console=tty1 console=ttyS1,115200" to the linux kernel arguments. Here the serial console over ipmi will be the default during the install process, while the other console should echo the output of the ipmi console.

After install the kernel argument "console=ttyS1,115200" was already added to the kernel boot arguments. I have additionally added "console=tty1" before this, this may be required to enable interaction with the server via a directly connected terminal if needed.

With the ipmi port set as default (last console specified in the kernel arguments) SL6 will automatically start a getty for ttyS1. If it was not the default console we would have to add a upstart config file in /etc/init/. Note SL6 uses upstart, previous SL5 console configurations in /etc/inittab are ignored!

e.g. ttyS1.conf

start on stopping rc runlevel [345]
stop on starting runlevel [S016]

respawn
exec /sbin/agetty /dev/ttyS1 115200 vt100

Software Sustainability Institute Blog | Software Carpentry bootcamp in Krakow

2013-05-31T09:40:11+00:00

By Aleksandra Pawlik.

This May, Software Carpentry once again went European with Jagiellonian University hosting the first bootcamp in Poland. There were 28 attendees: most of them postgraduate students and a few faculty members. They represented a range of disciplines from mathematics and theoretical physics to biology, genetics and medicine.

Software Sustainability Institute Blog | Do you know who owns your Intellectual Property?

2013-05-30T09:25:11+00:00

By Simon Hettrick.

Intellectual Property (IP) is a thorny issue in academia, generally because people don't know who owns what. There's a lot of hearsay, rumour and - frankly - misunderstanding about IP ownership, so we've decided to run an experiment. And we're going to need some guinea pigs...

The provisions that govern IP ownership appear fairly straightforward. Your IP is your own unless you work in a role where you're expected to generate IP, in which case your employer owns it.

Things get more complicated, when people start to ask "Are PhD students employees?" and "Do I own the software if I wrote it on my work computer, but outside of working hours?". Add to this employment contracts that deal with the assignment of IP, even though these often go unsigned and some contain clauses that are not legal. And, of course, the fact that who owns what is often not discussed... right up until someone creates something valuable and tries to make money from it.

Software Sustainability Institute Blog | Ontologies…handle with care!

2013-05-28T09:10:37+00:00

By Mike Jackson.

A couple of years ago I was involved in a project for which we needed an ontology. I was very uneasy as prior experience, over the years, had taught me that ontology meant meetings that overran and the same discussions, with no outcomes, over and over, a conceptual Groundhog Day. But my worst fears were, thankfully, not realised…

The project I was involved in, SPQR, was concerned with epigraphy, information about inscriptions – such as grave markers, tablets, signs or graffiti – from the ancient world. Epigraphic data is often fuzzy and incomplete due to uncertainty over where and when inscriptions originate and to who they refer. SPQR’s goal was to see whether using linked data to represent epigraphic data and join multiple data sets together offered a means by which the combined data could be more readily explored and browsed than relational or XML representations traditionally used.

Software Sustainability Institute Blog | Top tips for porting an application to a computing infrastructure

2013-05-22T17:29:50+00:00

By Mike Jackson.

You have access to a new computing infrastructure – a cluster, grid, cloud, HPC facility or volunteer computing infrastructure – and you wish to get your application running on it. This may be as simple as copying across an executable and running it, or, it may equally require a significant re-working of your application’s configuration or code. How you proceed depends upon your preferred working practices – you may want to scope out the work in advance in some detail or you may just want to get on and experiment.

How you port your application depends upon your application, your infrastructure and your personal preferences. However, there are a number of things that have the potential to make the porting go smoothly and that will also help others who wish to undertake similar work in the future. Here are our top tips for porting your application…

Software Sustainability Institute Blog | Top ten reasons to not share your code (and why you should anyway)

2013-05-20T07:11:59+00:00

By Randall J. LeVeque, professor in the Department of Applied Mathematics at the University of Washington in Seattle.

This post was originally published as a news item on the Society of Industrial and Applied Mathematics website (SIAM News, Vol. 46, April 2013).

There is no... mathematician so expert in his science, as to place entire confidence in any truth immediately upon his discovery of it... Every time he runs over his proofs, his confidence encreases; but still more by the approbation of his friends; and is raised to its utmost perfection by the universal assent and applauses of the learned world.

---David Hume, 1739*

I am an advocate of sharing the computer code used to produce tables or figures appearing in mathematical and scientific publications, particularly when the results produced by the code are an integral part of the research being presented. I'm not alone, and in fact the number of people thinking this way seems to be rapidly increasing, see for example [1–3, 6–8, 10].

Software Sustainability Institute Blog | The invaluable contribution of open-source software to research

2013-05-17T08:01:14+00:00

By Mark Woodbridge, software developer, Bioinformatics Support Service, Imperial College London.

This is the first article in a new series called a day in the software life. In this series, we will be asking researchers from all disciplines to discuss the tools that make their research possible.

I'm a research software engineer developing tools to help life scientists organise, analyse and share their data. This is a varied and rewarding role but with its challenges - not least keeping up to date with both technology and the relevant research whilst trying to build robust, usable, self-sustaining software and getting credit for doing so. However, one thing that makes it enjoyable is being part of a wider development community. The innumerable open source software projects that enable or accelerate development of the tools that we build for others are an invaluable resource and were the basis of a recent discussion at Collaborations Workshop 2013.

Software Sustainability Institute Blog | Top tips for choosing a computing infrastructure

2013-05-16T13:22:15+00:00

By Mike Jackson.

Deciding upon a computing infrastructure for your project can be a daunting task. Do you want to use your institution’s existing infrastructure, buy a new one, or go for a third-party? Would an in-house cluster be best or a commercial cloud platform? Regardless of whether you seek a cluster, grid, cloud, HPC facility or volunteer computing infrastructure, there are a number of things you should consider.

Here are our top tips for choosing a computing infrastructure…

Software Sustainability Institute Blog | The Oxford Doctoral Training Centre gets Software Carpentry

2013-05-15T07:08:45+00:00

By Mario Antonioletti.

A version of this post originally appeared on the EPCC blog.

I participated as an instructor in a Software Carpentry bootcamp that took place this week in Oxford. The bootcamp was organised by Jonathan Cooper and targeted at researchers involved in the Oxford Doctoral Training Centre. Shoaib Sufi from the Institute was the other instructor at this event. The three of us taught about 30 attendees from various disciplines studying for DPhils (this being Oxford) as well as some Postdocs, giving them some basic computing skills that we hope will make their research more productive.

Software Sustainability Institute Blog | If we value software, we need "a fundamental sea-change" in academia

2013-05-14T11:02:14+00:00

By Alice Tobin, science writer.

Combining the skills of a scientific researcher and a software developer, the research software engineer is ideally placed to bring scientific software up to scratch. An ongoing discussion that began at the Collaborations Workshop 2012 asks what obstacles need to be removed to clear the way.

Beneath the arched ceilings and elaborate chandeliers of Queen’s College, at the University of Oxford, an odd mix of researchers and software developers congregated last March. The 2012 Collaborations Workshop, organised by the Software Sustainability Institute (the Institute), brought together people from all sorts of backgrounds to discuss the future of scientific software development. And it was here, under the gaze of beautifully robed portraits, that they began to map out a new career path in the difficult terrain of academia.

Software Sustainability Institute Blog | Top tips for packaging software

2013-05-13T15:29:17+00:00

By Mike Jackson.

The first encounter a user often has with your software is when they download it as a package. The nature of this package, and what the users have to do with it, can make the difference between a happy user using your software for their work, and a frustrated user, who has wasted a morning, and is now looking elsewhere.

Here are our top tips for packaging software…

1. If you can, distribute a binary

If you can distribute a binary executable, your users will thank you. This might be an EXE file for Windows, or an executable JAR file, for example. But, if all a user has to do is double-click on a file or run one or two commands - whether that be a “yum install”, “easy_install” or “java -jar application.jar” - to install or run your software, then they’ll be off and running far more rapidly than if they first have to try to build it.

Software Sustainability Institute Blog | Cutting edge tools for cutting edge results

2013-05-07T13:34:55+00:00

By Kayla Iacovino, University of Cambridge

Remember research before computers? I sure don’t. But I’m told that back then, people still used slides to give talks. Figures were hand drawn, and references were found after hours upon hours in a brick and mortar library going through stacks of paper journals.

These days, we’ve automated a lot of the more mundane tasks involved in doing science. As someone doing their PhD in 2013, I’ve never had it any other way.

Even among us geologists, where one might suspect the research techniques to be as old as the dirt we study, science moves at a high-tech pace. One of the common tools used to analyse rocks is a type of vibrational spectroscopy known as Fourier Transform Infrared (FTIR).

Northgrid-tech | A thing of beauty Digital R81

2013-05-03T15:38:28+00:00

We needed to make space for a couple of new racks and decided to get rid of our 'workbench'. The workbench consisted of an old Digital VAX system which was mostly stripped of it's innards to leave a few sturdy steel frames. Here are some nostalgic pics of the last remaining unit being gutted. Note the gorgeous circuit boards so easily accessible. And yes, that last one is a 500MB harddrive (so I'm told). The motor seems more suited to a washing machine. They don't make 'em like they used to.

Software Sustainability Institute Blog | Why are so many biologists scared of programming?

2013-05-01T11:15:14+00:00

By Jane Charlesworth, University of Oxford.

I am a biologist. I also love computers. Sadly, I have many colleages who find programming scary, and I used to be one of those people. Here is how I learned to stop worrying and embrace programming as a tool for my research.

All the tech skills are self-taught, or picked up from eavesdropping on computer scientist friends in the pub. Most of my colleagues are either in the same boat or reluctant to spend their research time learning to write reproducible code. Even those of us who are eager to learn don't know where to begin. I think this represents a crisis in biology education.

I did a degree at a top UK university about ten years ago. A good 95% of my degree consisted of training for working in a wet-lab, doing molecular biology. I have never worked in a wet-lab. Granted, when I did my degree, the first human genome was still unpublished, but I find it impossible to believe that our course organisers could not have foreseen that sequencing was a rapidly maturing technology and the students they were training would need to be equipped with the skills to deal with the resulting deluge of data.

GridPP storage news | My new sister Eve

2013-04-26T06:13:15+00:00

SO I have a new friend Eve. Eve is my last sister born in 2012 whose first home is the same as mine.
Initial info of Eve to compare with mine and Georgina follows. What surprises me is the number of files that that have no replicas at all and so are at risk if a house or room gets destroyed.

DataSet Name			Dave	Georgina	Eve
"DNA" Number			2.3.7.3623	3.3.3.3.5.5.89	3.53.1361
Number of Countries			17	17
Number of "Houses"			59	79	63
Type of Rooms:DATADISK			28	49	50
Type of Rooms:LGD			37	61	14
Type of Rooms:PERF+PHYS			32	58	22
Type of Rooms:TAPE			6	12	9
Type of Rooms:USERDISK			1	9	5
Type of Rooms:CERN			3	6	5
Type of Rooms:SCRATCH			0	12	14
Type of Rooms:CALIB			0	5	7
Total number of people (including clones)			1166	1594	642
Number of unique people			894	1120	299
Numer of "people" of type:		^user	137	470	64
Numer of unique "people" of type:		^user	132	429	55
Numer of "people" of type:		^data	725	1048	538
Numer of unique "people" of type:		^data	532	631	212
Numer of "people" of type:		^group	31	70	40
Numer of unique "people" of type:		^group	26	54	32
Numer of "people" of type:		^valid	1	0	0
Numer of unique "people" of type:		^valid	1	0	0
Datasets that have	1	copy	691	811	178
Datasets that have	2	copies	143	197	63
Datasets that have	3	copies	52	68	25
Datasets that have	4	copies	71	23	7
Datasets that have	5	copies	1	7	21
Datasets that have	6	copies	0	5	2
Datasets that have	7	copies	0	2	1
Datasets that have	8	copies	0	1	1
Datasets that have	12	copies	0	0	6
Datasets that have	13	copies	0	0	3
Number of files that have	1	copy	53644	141663	9509
Number of files that have	2	copies	12904	36863	7705
Number of files that have	3	copies	1860	3708	4520
Number of files that have	4	copies	7	2092	110
Number of files that have	5	copies	7	572	1091
Number of files that have	6	copies	0	487	20
Number of files that have	7	copies	0	142	5
Number of files that have	8	copies	0	73	1
Number of files that have	12	copies	0	0	6
Number of files that have	13	copies	0	0	5
Total number of files on the grid			85095	242241	22973
Total number of unique files on the grid			68422	185600	44674
Data Volume (TB) on the grid that has	1	copy	8.8	27.7	1.7
Data Volume (TB) on the grid that has	2	copies	6.8	21.3	4.7
Data Volume (TB) on the grid that has	3	copies	0.3	5.6	7.2
Data Volume (TB) on the grid that has	4	copies	0.011	1.5	0.12
Data Volume (TB) on the grid that has	5	copies	0.028	1.4	0.47
Data Volume (TB) on the grid that has	6	copies	0	0.36	0.039
Data Volume (TB) on the grid that has	7	copies	0	0.13	< 1GB
Data Volume (TB) on the grid that has	8	copies	0	0.12	< 1GB
Data Volume (TB) on the grid that has	12	copies	0	0	< 1GB
Data Volume (TB) on the grid that has	13	copies	0	0	< 1GB
Total Volume of data on the grid (TB)			23.7	104	35.8
Total Volume of unique data on the grid (TB)			16	58	14.2

Software Sustainability Institute Blog | Rewarding researchers that "strive for reproducibility" in audio and music research

2013-04-25T14:04:12+00:00

By Luis Figueira, SoundSoftware

We are very excited to announce the first competition for SoundSoftware.ac.uk prizes for reproducibility in audio and music research. Our goal is to promote the development and release of sustainable and reusable software and datasets alongside published research.

In a few of our recent papers and presentations we've seen that much audio and music research work is published without the accompanying software implementations. One reason is a lack of self-confidence - the fear that one's code is not good enough to share. (You can read more about this in our ICASSP 2012 paper, Towards Software Reuse in Audio and Music Research and in the Institute's post Haters Gonna Hate - why you shouldn't be ashamed of releasing your code). We believe that one way of countering this fear is to promote the idea that sharing is a worthwhile goal in itself.

Software Sustainability Institute Blog | A Digital (R)evolution in Palaeontology

2013-04-23T15:35:28+00:00

By Stephan Lautenschlager, Fellow and PhD Candidate at the University of Bristol

Palaeontology, the study of fossils and extinct organisms, has a reputation for being a very traditional and dusty discipline. Partly influenced by Palaeontologists’ portrayal in the media, the public imagines them roaming the badlands in the search for new fossils. Alternatively, they are seen as ivory tower scientists, spending most of their time in museum collections studying the bones of long extinct animals. Although both images hold a partial truth, palaeontology has experienced a tremendous paradigm shift in the last decade and has evolved into a multi-disciplinary science. Advances in hardware and software, and their wide and comparably inexpensive availability have led to a surge of computational methods to study fossils.

Software Sustainability Institute Blog | Introducing three great open-source packages

2013-04-23T14:00:58+00:00

By Alexander Hay.

What is the best open-source software? This is a question I decided to answer, and so began a long and no doubt eventful journey of discovery. One of my first destinations was at an open-source break out session, which took place last month at the Collaborations Workshop 2013. Cue much debate.

It's not possible (and probably not helpful) to decide on the best software overall, so instead I have focussed on open-source software that is exemplary in certain areas. Here follows the first three examples of this software and why we chose them. In future posts, I will work through the other examples on our list.

Software Sustainability Institute Blog | Don't forget the people - a fractured training landscape

2013-04-22T08:00:17+00:00

By Simon Hettrick.

This article about the EGI Community Forum was originally posted on the Gridcast blog.

In his keynote address at the EGI Community Forum, Peter Coveney described the need to combat the fragmentation of e-Infrastructure. Unlike other talks I have seen on this subject, Peter’s focus went beyond the hardware "it’s not just the tin and iron boxes but the software and - most importantly - the people". It is the human capital, the well informed and correctly trained researchers, that we need to make the most of e-Infrastructure.

LondonGrid | The art of cabling

2013-04-21T23:43:13+00:00

The challenge of organising your cables behind your TV is nothing compared to that of a large computing cluster.

One of our standard racks contains 12 Dell R510s servers (for storage) and 6 Dell C6100 chases (providing 24 compute nodes) all 36 nodes are connected with a 10 Gb (SFP+), 1 Gb (backup) and 100 Mb (for IPMI) network cable. Connecting to 3 different network switches at the top of the rack. In addition the 18 "boxes" need a total of 36 power connections. A total of 144 cables per rack!

How to cope? Separate the network cables from the power cables, a possible source of noise. Use different colour cables for the different traffic and add unique id number for each cable. Use lose, removable cable ties. When a cable brakes don't remove it, just add a new cable.

The 10 Gb switches, in our case Dell S4810s, connect using 4 40Gb QSFP+ cables to two Dell Z9000 core switches. Having two core switches allows us to take one unit out of service without downtime (we use the VLT protocol and it works!). However this does add cable complexity. The backup 1 gig switches connect to each other in a daisy chain using 10 Gb cx4 cables, left over from before our 10 Gb upgrade. Finally the ipmi switches connect to a front-end switch using 1GBaseT cables.

The picture shows the inter-switch links. Visible are the orange 40Gb connections and blue 10Gb cx4 cables. In addition each 40 Gb cable has an ID indicating which rack it came from and which core switch its going too.

We have one rack full of critical, world facing servers. These servers need to be available all the times making it very difficult to reorganise the cabling. As a result over time, as we add and remove servers, the cabling becomes a mess. This is starting to become a risk! We are just going to have to accept some down time to sort it out in the near future.

Software Sustainability Institute Blog | Software Carpentry makes its debut in Manchester

2013-04-19T18:21:05+00:00

By Mike Jackson.

After last week's EGI Forum, I returned to The University of Manchester where I led a Software Carpentry boot camp alongside my Institute colleague Aleksandra Pawlik, and David Jones from the Climate Code Foundation. Both Aleksandra and David, who helped out at the Edinburgh boot camp in December, made their instructor debuts.

Software Sustainability Institute Blog | Software Carpentry highlights from the EGI Community Forum

2013-04-16T12:20:34+00:00

By Mike Jackson.

At last week's EGI Community Forum, we hosted a day of Software Carpentry boot camp highlights. These were taster sessions drawn from Software Carpentry's highly successful boot camps, which are for researchers who want to learn software development skills to do more in less time, and with less pain.

Using version control to record provenance and collaborate more easily.
Using testing to help ensure your software, and results, are correct.
Data management using a NoSQL database to manage your data more easily.

LondonGrid | virtualization performance hit

2013-04-15T11:32:02+00:00

Like the rest of the world, there is a lot of discussion going about the use of clouds and virtualization in gridpp.

http://www.gridpp.ac.uk/gridpp30/mcnab-lhcb-vmclouds-march-2013.pdf

http://www.admin-magazine.com/HPC/articles/the_cloud_s_role_in_hpc

Using virtualization will have a performance impact, so using it for our type of computing (hpc/htc) may not be the best solution. However just what impact does it have? A quick search of the web suggests anywhere between 3 to 30%. Most of the overhead appears to be in the kernel and in i/o.

http://serverfault.com/questions/261974/how-much-overhead-does-x86-x64-virtualization-have

http://www.altechnative.net/2012/08/04/virtual-performance-part-1-vmware/

http://www.anandtech.com/show/3827/virtualization-ask-the-experts-1

I decided that I wanted to do some of my own tests with the focus on the type of work we do in gridpp.

Testbed: 24 thread westmere processor running at 2.66 GHz + 48 Gig of memory using Scientific Linux 6.3 (basically RHEL6). I'm using the default install of KVM with the virtual image as a local file setup to use all 24 threads.

Benchmarks: 1) I unpack and make the ROOT analysis package using 24 threads; 2) as 1 but using only one thread. 3) I generate 500,000 Montecarlo events using the HERWIG++ Generator; 4) as 3 but I also include the time taken to unpack and install HERWIG++; 5) I run the HEP-SPEC06 benchmark. For tests 1 to 4 i use the TIME command to obtain the real time taken (smaller is better), for 5 I report the hep-spec score (larger is better). I will run the benchmarks on the bare metal install and on the VM on the same hardware and compare the results.

Results:

Out of the box performance of KVM results in ~3% (CPU intensive) to 20% (sys call intensive) reduction in performance. There is some indication of correlation with ratio of sys time / user time (particular effect with make/tar/gzip?). This is not seen in HEP-SPEC result. SYS time is the CPU time spent within the kernel and from previous studies we expect this to incur a high performance hit in virtualization.

If I get the time I intend to repeat analysis using optimisations (e.g. guest image on LVM). Repeat analysis using fedora 18 ( ~RHEL 7). Repeat using sandybridge cpu. Look at network performance (eg iozone with lustre).