tcpip: (Default)
It's been a very full week of a working holiday in Townsville teaching high performance computing, parallel programming, and the like at the Australian Institute of Marine Science. There were about fifty or so researchers in attendance, mostly in person and a few online. Although we had an official programme that ran from 9 to 5, out-of-hours discussions pretty much took up the rest of the time; mostly on HPC work, marine science, climatology, and languages. These were mostly youthful and very smart and a switched-on group who were very engaged with the content throughout the week, and very keen on getting assistance for their specific problems. It was quite delightful watching the (relatively small) cluster rocket in usage over the week as soon as they were taught how to fine-tune and submit various parts of their job submissions, and I get the sense that although many are part of different research groups, they are all keen to build a community of HPC users. It was also great, I must mention, to spend time in the company of Geoff M., the tireless AIMS sysadmin whom I met there almost ten years ago, and Patrick L., a marine researcher who was also there on my first visit. Diego B., provided a powerful and driving force in getting this week organised and finally, I was quite charmed to discover that among the several Perth AIMS visitors, one Barbara R., was also at Murdoch University at the same time as myself and part of the SF club I started, MARS.

Buried in the institute's buildings I didn't get much of a chance to explore even the impressive natural surroundings of AIMS, although I did receive regular visits from the local wallabies and brush turkeys, and my on-site accomodation had more than a few geckos. A particularly large trapdoor spider was also in the vicinity which attracted a bit of attention. On the last day we finished early, so I was given a bit of a tour of some key parts of the facilities, including the "National Sea Simulator", the experimental coral reef growth labs and, of course, their data centre. No matter how often I visit such places, it never ceases to amaze me how relatively modestly such scientific research facilities are funded, how modestly the actual researchers themselves live, and how every spare dollar is spent not on keeping such facilities prettified, but rather on the science itself. I have nothing but complete respect for these people.
tcpip: (Default)
It has been a diverse week of the scientific nature for me which, unusually, covered not just work and study events but social events as well. Last night I was pleased to meet one Juliette W., with whom we share an interest in science fiction and fluid dynamics. We have mutual friends, but a surprising connection was with Robert McLay of the Texas Advanced Computing Centre - I attended his final pre-retirement presentation on LMod a few evenings prior (at 1 am in the morning, Melbourne time). LMod really has transformed the management of applications in the high-performance computing space over the past several years, so I hope Rob receives some recognition for his contribution.

Another event that crossed the scientific and the social occurred the night before when I caught up with a former workmate, Martin P., to see Oppenheimer at iMAX to a packed audience. It's a great topic and the representation of the science wasn't terrible even if the history was not quite right and it managed to omit the fallout effects. It must be said the performance by all and sundry was of excellent quality, and the story delved deeply into the persecution and troubled individual psychology of the protagonist. Ultimately, however, I was rather underwhelmed by it all - the movie was too long and it really didn't gain anything by being on a big screen. Still, small mercies that it wasn't a travesty of an action film.

In more typical activities, earlier in the week I had the pleasing opportunity to chair a work forum where we were addressed by two excellent researchers (Dr. Shanaka Kristombu Baduge and Dr. Sadeep Thilakarathna) who have usd our kit (cloud and HPC) to develop AI robots for plastic waste selection. In regards to the more typical study project, in the final lecture of the week on paleoclimatology and current events I drew attention to recent warnings of a very probable collapse of the Atlantic Meridional Overturning Circulation and raised whether it was possible to have a collapse in North Atlantic ocean temperatures whilst having simultaneously a significant increase in atmospheric temperatures. It seemed plausible and but I was hoping Professor Rewi Newnham would answer in the negative; he didn't. Fun time ahead, eh?
tcpip: (Default)
As 2022 comes to a close once again the utility of scientific research becomes evident, through incremental progress and the occasional breakthrough. One such big breakthrough in recent days was successful fusion ignition at the National Ignition Facility (NIF) at Lawrence Livermore National Laboratory; more power was produced than what was put in. True, it was a very minor proof, and there was significant energy put into the lasers to begin with, but it should not be understated - fusion energy would provide abundant, cheap, and safe energy - if the hard limiting issues of temperature, pressure, and duration can be resolved. This success is much more significant than, say, the successful "artificial sun" from China's Experimental Advanced Superconducting Tokamak in 2018, or the generator from the first stellarator at Max Planck Institute for Plasma Physics in 2015.

This year has also witnessed a succession of interesting discoveries and inventions in cancer research. The application of dostarlimab, used to treat endometrial cancer, resulted in all 12 people being treated at the Memorial Sloan Kettering Cancer Center (MSKCC) with rectal cancer tumors being cured. Although this is a small trial, it is something that has never happened in the history of cancer. Also, the University of Texas Dallas synthesized a novel compound called ERX-41 that has shown promise in eliminating a broad spectrum of hard-to-treat cancer cells, including triple-negative breast cancer. Just a few days ago, a Moderna mRNA melanoma vaccine with personalised treatment reduces recurrence or death by 44% compared to
immunotherapy drug alone - the cookers won't like that one.

To finish the trilogy, one cannot help but notice the explosion of interest in AI over the past few weeks in social media. In part, this has been due to the excitement and concerns around AI-generated art. For many consumers, it has been a toy for many artists they have become nervous, and my favourite gaming company, Chaosium, has released a statement banning AI-generated art in their books. I am not convinced by their reasoning, but I certainly understand the motivation. As an alternative, I am interested in comparison that various programming forums are largely enthusiastic with tools like OpenAI in producing fairly reasonable code, and an explanation, all faster than StackOverflow, for example. I wonder whether the "procedural music" being produced by 65dos (No Man's Sky, Wreckage Systems) counts? Either way, it makes for an interesting space to watch, and I will stick to the argument that inevitably anything that can be automated will be automated. This does mean a challenge to the value of labour and the ownership of capital, and it will mean a challenge to for human beings in their own inventiveness, creativity, etc.
tcpip: (Default)
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".

Profile

tcpip: (Default)
Diary of a B+ Grade Polymath

October 2025

S M T W T F S
   1 23 4
567891011
12131415161718
19202122232425
262728293031 

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Oct. 5th, 2025 01:08 am
Powered by Dreamwidth Studios