Lev Lafayette's journal

Every year I do a series of invited lectures and assessment for the UniMelb master's-level course "Cluster and Cloud Computing", which is a rather massive course with approximately four hundred students primarily from computer science and data science. Anyway, this has started for 2025 and will continue for the next few weeks. In between all this I have two advanced researcher workshops next week on "Parallel Processing" (shell commands, job arrays, OpenMP and OpenMPI programming), and "Mathematical Programming" (R, Octave/MATLAB, Julia, Maxima, Stata, etc). Further, I recently gave a presentation to the Spartan Champions group on scheduler-level parallelisation. In most cases this is the easiest way to do a type of parallel data computation and, as such, makes a low-hanging opportunity for making the best use of computing resources. Finally, on a slightly related matter, I have put up a script (called "lament") designed for local viewing of Facebook encrypted messages, converting the JSON to HTML conversion of Facebook Messenger files using the Pandas and JSON extensions for Python.

I have had the pleasure of catching up with interstate visitors this week! Firstly, I had the delight to catch up with Justine, Simon, Erica, and Susie C., for brunch with "Le Cafe Flo, a rather good French cafe in Thornbury. It was conveniently located to "The Witches Wardrobe Flea Market", which had a good variety of goth, emo, etc. wares. It was a bit of a Perth migrants day for our part, enhanced with visitor Susie C., who is making a lot of noises about making the trip across the Nullabor on a more permanent basis. Another visitor this week is my dear friend (and co-owner of property) from Darwin, Lara D. Lara is blessing me with two weeks with her company, and we have quite the range of artistic, musical, and culinary delights lined up, although it will be hard for the latter to beat the dinner provided by Carol D., Lara's mother, last night.

I really feel like I've outdone myself this week, looking at the range and quantity of activities. Of note includes two meetings of the National Tertiary Education Union, one on the future of public education and the budget (with a very good guest speaker, Prof Dr Julian Garrizmann, from the Goethe University Frankfurt, who was steadfastly evidence-based), and the second on Gaza and the student encampments (three motions passed with 1% opposed, demanding disclosure of UniMelb's links to the war, supporting academic freedom to speak on the issue etc). Matters on the former perspective were followed up with a long night with the local ALP branch on the Federal budget and then continuing discussions well into the night with several of the younger branch members. Most people think the budget was pretty good, and it was, with the glaring exception of that enormous bugbear expense, the AUKUS submarines, whose scale I believe is incomprehensible to most.

On Sunday a number of people (myself, Rodney, Andrew, Charmaine, and Penny) from the RPG Review Cooperative ventured out to the outer suburbs to visit our friend Michael who is currently in convalescence. He is in good spirits and seems to recovering well, but tiredness is a factor. In related news, Erica and I visited David and Angela recently who are in the midst of moving from their long-term abode. David kindly gifted us a veritable mountain of roleplaying games and comics (about two bookshelves worth), and Erica and I took great wide-eyed delight fossicking through them over dinner. Speaking of such things, Friday night's south-western Chinese feast with Liana, Julie, and James was an absolute joy and we even managed to finish a game of Trivial Pursuit, which has dated quite significantly. Of another culinary adventure, Ruby's visit came with the impromptu invention of a (spinach and blue-cheese based) béchamel senfsauce verte, in my apparently never-ended attempt to combine French and German cuisines.

All play and no work is, of course, implausible and whilst most of my work is invariably going through some difficult optimised scientific software installs, I have been very impressed with one recent workflow that involves the automatic generation of array job scripts which themselves have job dependencies, which in turn call another set of job arrays. The fact that there have been several support requests this week that are of unusual levels of complexity has been challenging and rewarding, as has a review of the workplace's "five-year plan" with which I hope I have made some reasonable suggestions for clarification, elaboration, and improvement.

With time running out for continuing registration, I have quickly put together an annual general meeting for the Isocracy Network next Saturday with yours truly speaking on "Climate Change and International Politics", which I believe I might know something about. I've scurried to get everything ready for that day, including the annual report, and getting formatting for articles on the website correct; the most recent being "The Case for Opposing the AUKUS Agreement" by Labor Against War, "Cash is an Anachronistic King" by yours truly, and "Reviewing 'At Work in the Ruins'" by Robert Barker. There is also one on the recent events in Gaza forthcoming and, of course, notes from Saturday's presentation will also be included.

For the past five days in succession, I have been involved in presentations in some manner. Today I attended the seminars for our paleoclimatology paper at the University of Wellington and I presented on "The Younger Dryas and The Atlantic Meridional Overturning Circulation" at Wellington University. The YD is clearly a topic of great interest as one other researcher gave a paper on the "impact hypothesis" as a critical review (it's almost certainly not true) and another in the context of volcanism (a contributing factor, but probably insufficient). Other seminar presentations included the idea of microplastics as a marker for the Anthropocene (I rate animal population extinctions and habitat change a little higher, but plastics can certainly be an "indicator species"), along with one on animal and plant population changes in the New Zealand Southern alpine regions due to climate change.

In addition to this, yesterday I chaired the presentation by Dr Matthew Burns and Tom Wilkins from the University's Waterway Ecosystem Research Group, who spoke about the implementation of "backyard catchments" and invertebrate populations. Preceding this was three days of Spartan Upgrade workshops that I conducted; the second day was probably the most interesting with the most relevant questions on the new software stack. Next week there is another presentation at Wellington University, but that is a group project (climate change mitigation in the Te Tairawhiti region - fortunately someone else has stepped up to do the actual presentation, so I'll just be making contributions to that.

It is probably opportune to mention in this context that I'll be visiting China is a couple of weeks, specifically Wuxi for the "International Conference on Green and Innovation-driven Development in Cities and Towns". I attended the first conference last year (online) and was quite impressed by the emphasis on presentations from an engineering, town-planning, and ecological background, and of course, the city itself has been part of a low-emissions pilot program for some years. This will be my first visit to said country and, as can be expected, the visa application was quite a hefty form, but all processed quite efficiently.

This evening I completed a difficult short exam for "Physical Basis of Climate Change", just the final exam to go. I also received a grade today for the major essay for the same subject on "Earth's Climate System and Changes Since the Industrial Revolution"; given that I received 27/30 maybe I know something about this. Given that this is one of the more challenging subjects I have taken in my long and varied university career, I am understandably quite pleased with this result. The next few weeks will see final assessments in this and other subjects, and that will be trimester one complete for this degree.

There's been a couple of nice upticks at work today as well. This afternoon I hosted a researcher presentation for one Arshiya Sangchooli who gave a great talk about amygdala and information processing using Spartan and Mediaflux. Apart from speaking on a part of the brain that holds a particular interest to me, it is always great to see how very complex problems that require a lot of data are processed on our system to generate useful results. In addition, a survey of staff from the Cultural Working Group suggested that my work for the past two-plus years in this body has not been in vain, with very significant improvements across all previous metrics of concern. More work to be done, but it was a very pleasing result.

Tomorrow is the annual general meeting of the Isocracy Network, my favourite political organisation (it should be, I founded it). We're having a discussion on recent increases in rents, housing prices, interest rates and the like and why home affordability has become increasingly painful for many Australians. It is a subject that I've been grumpy about for some years but - rather like global warming - there are some powerful vested interests that get in the way of making life better for people. As a related political aside I must mention attending a Melbourne adieu for one Doone Clifton who is moving interstate. Doone is an old North Melbourne Labor Party comrade who I first met over twenty years ago, and her farewell really was quite a meeting of people of that locale and politics. It was also a lovely opportunity to see Rob and Angela L., there as well with follow-up drinks and conversation with these worthy souls.

I rather wish there was an additional week this month. That would at least give me the opportunity to voice my rather strong opinions about the ludicrous submarine purchases, obselete before they hit the water and the opportunity costs that come with them. I would also have a few, more beneficial, words to say about the Voice for First Nations people referendum, and earlier this week I attended a forum at the St Kilda Cricket Club with Senator Jana Stewart speaking, along with Josh Burns (Federal MP) and Nina Taylor (State MP) in attendance. Comments on these matters will come soon; I also have to organise an Isocracy AGM for next month on the topic of housing affordability; there are reasons for high rents, mortgage stress, and higher interest rates.

Instead, I have been buried in teaching, learning, and work. This week I ran two Linux and HPC workshops and next week I have tutorials and lectures to deliver for the University's Cluster and Cloud Computing Course. Among these, I've had some particularly tricky software installations to do, involving far too many R packages in two cases and a recent release of PyTorch (have you tried building that from source? So many dependencies, so many patch files needed). Among all this, I have buried myself in the coursework for my last Psych unit and the four Climate Science units, along with finishing the Yale University "Science of Well-Being Course" (it made very little difference to me, it would make a big difference to others). True, I am ahead in everything in this regard, but I need to be. This is one of the maddest months in recent years in terms of workload.

Still, there have been a couple of social occasions. Earlier in the week I was visited by John August, National Treasurer of the Pirate Party, who was briefly visiting from Sydney and we conversed about various matters of philosophy, politics, and some "difficult" people. Last night I had yet another visit to The Capitol Theatre (it's becoming my second home) with Erica H., to see a screening of the Hitchcock classic "Psycho", which come with an introductory lecture by no less than Professor Adam Lowenstein of Columbia University, who is on tour. Erica is probably the biggest Hitchcock fan I know, so the opportunity couldn't be missed for this rather brilliant piece of film art, and Lowenstein's lecture provided additional insight.

Most people use their computers (which includes mobile phones) for communication, social media, games, entertainment, office applications, and the like. Most of the time these activities are not particularly onerous in terms of computing as such or do not lead to enormous benefits in productivity, inventions, and discovery. There is one field, however, rarely discussed, that does do this - and that is supercomputing. It is through supercomputing that we are witnessing the most important technological advances of our day, including astronomy, weather and climate forecasting, materials science and engineering, molecular modeling, genomics, neurology, geoscience, and finance - all with numerous success stories.

Usually, I draw a distinction between supercomputing and high-performance computing. Specifically, a supercomputer is any computer system that has exceptional computational power at a particular point in time, many (but not all) of which are measured in the bi-annual Top500 list. Once upon a time dominated by monolithic mainframes supercomputers, in a contemporary sense, are a subset of high-performance computing, which is typically arranged as a cluster of commodity-grade servers with a high-speed interconnect and message-passing software that allows the entire unit to be treated as a whole. One can even put together a "supercomputer" from Raspberry Pi systems, as the University of Southhampton illustrates.

How important is this? For many years now we've known that there is a strong association between research output and access to such systems. Macroeconomic analysis shows that for every dollar invested in supercomputing, there is a return of forty-four dollars in profits or cost-savings. Both these metrics are almost certainly going to increase in time; datasets and problem complexity are growing at a rate greater than the computational performance of personal systems. More researchers need access to supercomputers.

However, researchers do require training to use such systems. The environment, the interface, the use of schedulers on a shared system, the location of data, is all something that needs to be learned. This is a big part of my life; in the last week, I spent three days teaching researchers from the basic of using a supercomputer system to scripting jobs, to using Australia's most powerful system Gadi at NCI, along with contributions at a board meeting of the international HPC Certification Forum. It is often a challenging vocation, but I feel confident that it is making a real difference to our shared lives. For that, I am very grateful.

On rare occasions, I will write about my work week and this is one of those times. As the previous installment revealed, this week I conducted three training workshops, "Introduction to Linux and HPC", "Advanced Linux and Shell Scripting for HPC", and "Parallel and High Performance Python". All of this was conducted whilst fighting off an annoying chest-and-head cold. As a result, I don't think I delivered at 100% but fortunately, the researchers seemed to be satisfied with the delivery and content. The Python course was a new workshop and I feel that it will have several more iterations before it is up to the standard I want, but even at this stage, there is a temptation to do a follow-up workshop that's even more advanced. There is much that I can write about this often slow and annoying (but readable and popular) language, but as a taste of things to come I have written a short article entitled "Compiling Your Python", where I take up a matter that I've been meaning to write about for quite a while.

Yesterday, I tried my best to take a day of sick leave to alleviate this cold but alas, I remembered almost at the last minute that I was chairing a researcher presentation, so at the last minute managed to get myself together. The presentation itself, by one Professor Guillermo Narsilio was excellent, as they all are. This one was about the development of various geothermal technologies, their capacity to mitigate climate change, their role in certain large infrastructure projects, and the modelling they did using the Spartan supercomputer and the University of Melbourne/NeCTAR Research Cloud. Finally, to provide a capstone to the week I helped out a researcher who was conducting a looped Bayesian phylogenetic and phylodynamic data integration for quality control but was having some leading character issues; the recommended solution of using a singleton scheduler directive to manage the outputs - and it worked! "Thanks for your advice too, Lev. You saved my PhD chapter!", Now that's the sort of feedback that gets me out of bed in the morning.

On Wednesday I went to an RMIT lecture and concert at The Capitol Theatre where alcohol was the musical instrument, i.e., some scientists from the ARC Exciton centre applied a nuclear magnetic resonance spectrometer to ethanol, producing a pitch in G-minor. The event was hosted by RMIT and hosted at their rather beautiful "Chicago-gothic" Capitol Theatre, which has been pretty much closed for the past two years. I went with Simon S., the second night in succession with his company, with he and Justine visiting the evening before for dinner and drinks before we went to the Ian Potter Centre for a performance from UniMelb's School of Fine Arts and Music. Before you all gasp "Lev went to a jazz performance!", I wish to point out that it was competently performed and subtle lounge music. To add to further events, this evening Mel and Vanessa S visited for dinner and drinks, and then we went to the New Music Studio "Of Birds and Monuments" by The Melbourne Conservatorium of Music which was heavily influenced by narration with the indigenous sites near the Melbourne CBD. Yes, three concerts in four days!

Apart from alcohol as an instrument, there has been a couple of interesting science-related matters that I've been involved in this week. One is a recent publication using the Spartan supercomputer which involved automated malware detection based on the image-based binary representation, with a pretty high level of accuracy. On another work-related science matter, I hosted a presentation today for Research Computing Services at UniMelb with Associate Professor Mohsen Talei on developing low-energy and cleaner gas turbines and reciprocating engines; it is from such engineering that a better planet can be made. Finally, in another work-related matter, I have spent a fair bit of time working on a high-performance and parallel Python workshop which I will run next month. Python is a great scripting language for beginners and has good object-orientated programming practices built in, which means that it is often used on an enterprise level as well. But if you care about performance and resource utilisation it's incredibly slow and inefficient. The workshop is designed to overcome some of these issues, but ultimately recommends "polyglot programming".

Every week there are one or two new research papers released that used the UniMelb supercomputer that I work on. This week's really caught my attention; "The impact of the COVID-19 pandemic on air pollution: A global assessment using machine learning techniques" published in Atmospheric Pollution Research. If there was a benefit of the pandemic (and they are few), one was the opportunity to quantify human activities on air quality; so modeling the effects in 700 (!) cities, they found that clean transport strategies in China and India would have the highest potential health benefit; whether or not anything happens will, of course, depend on public policy which is always a challenge. Today we managed to bring (almost) everyone into the office for a morning meeting summarising many of the utilisation and technology improvements we've had over the past year in high performance computing and use of the research cloud. I am increasingly of the opinion that given how broken our political economy is, which rewards the personality traits of narcissism, Machiavellianism, and psychopathy, the main vector to generating any improvements in people's wealth and well-being is select technologies.

Speaking of which, I am visiting Perth after an absence of over two years! I'll be departing on Thursday 9 June and returning Tuesday 14th June and dear West Australians, yes I will be arranging for another of my famous visiting dinners at some cheap and cheerful Italian restaurant, as is my habit on Sunday 12th. To say that I have missed the immediate company of my Perth friends over the past two years is quite an understatement. Despite being a Melbourne person for more than half my life now, the importance of those I associated with in those heady and exciting years and their development in the decades that followed is a critical part of my life. Of course, I haven't planned accommodation yet, but I do know that over the weekend I'll be meeting a couple of senior politicians, people from a certain embassy, and visiting a manufacturing plant. All of which is sufficiently ambiguous at this point, but I think I'll be able to reveal a lot more publically within the next fortnight or so.

There is now only sixteen days before the Australian Federal election, which has been relatively quiet. I suspect from this point onwards however campaigns are going to go ballistic. Whilst the cost of living is increasingly becoming the single most important issue, it is important to realise that a major driver of this issue is the direction of public monies. Monetary theory teaches us that a government in charge of its own money supply can basically create money ex nihilo (fun fact: taxes don't fund federal government expenditure like the NDIS, etc) and the real limiting factor is whether productivity gains equal the increase in the money supply. Which, apart from the sheer disgust at nepotistic behaviour, events like Aspen Medical receiving more than $1 billion in government PPE contracts despite their lack of large-scale procurement experience (but well connected to former LNP health minister Michael Wooldridge and current LNP minister Greg Hunt). It just horrifies me, and I believe this should be a sackable offense. Clearly, some people just don't care about good and careful stewardship of public monies, and it stands as another reason why we need a Federal Independent Commission Against Corruption. There are simply too many examples of very questionable behaviour.

I find myself quite concerned with public health policy and behaviour of the general public. As is my wont, with sober senses, I quietly compose my thoughts on matters, ensuring that I have brought together grounded evidence to generate considered convictions, this time comparing the public health policies of elimination from last year to this year's approach of vaccinated herd immunity: From Elimination to Herd Immunity?. Unfortunately, the reality is that herd immunity is a difficult number given the R0 of the Delta variant, and an elimination strategy is still the best, even if politically difficult, option. But the virus really doesn't care whether something is "politically difficult". There is a sense in which I personally feel a little surrounded by it all; there are no less than 22 coronavirus exposure times from 9 different sites surrounding me within 1km of where I live, including a block immediately to the rear of my apartment.

Despite this, I find myself engaging in some quasi-normal living, even attending a political webinar hosted by my local MLC, Nina Taylor, and addressed by the state minister of the environment, Lily D'Ambrosio, on the Victorian government's waste and recycling plans. A couple of days later I found myself visiting the St Kilda foreshore with Robbie K., who I count as a friend of some 25 years, although we haven't been in each other's company for the better part of more than 18 months or so. Once upon a time we would find ourselves attending various alternative concerts together (e.g., Jello Biafra, The Strokes, Massive Attack), now we wandered along the beach, reminisced, and grizzled about the effects of the pandemic and the huge quantities of maskless individuals.

I will also admit to some challenges at work in the past week as well. A quantum chemistry application, released as a collection of binaries only, has proven to be less than friendly with the version of OpenMPI that is linked to on our system (the serial version works just fine). A machine-learning image recognition application, designed for a different flavour of Linux and with a Makefile that dies when installing CUDA/GPU versions that don't exist, is also causing problems. Combining the two, I can also say the same for a molecular dynamics application, also designed for a different distribution but also with MPI errors. Still, I guess that having three work problems isn't too bad, but such problems! I am rather looking forward to delivering a class on Friday, which of course had all its places filled an hour after announcement. Apparently, people still want to learn about supercomputers.