## Archive for November 15th, 2007

### Supercomputer breakthrough from India

November 15, 2007

For an indigenously built supercomputer to make it to the fourth spot in the world ranking of high performance computing (HPC) systems is a real breakthrough. For this, the band of researchers at the Tata Group’s Pune-based Computational Research Laboratory (CRL), who achieved this just 18 months after their facility was established, deserve the highest praise. While the Tata system has been built around off-the-shelf technologies, the top three spots have gone to entirely bought out proprietary supercomputing systems, such as the IBM’s BlueGene. A comparison of the private effort with the record of the Centre for the Development of Advanced Computing (CDAC), a public-funded supercomputing research facility, is unavoidable. Although CDAC made the first Indian teraflop supercomputer, PARAM Padma, in 2003, its successor is yet to achieve the kind of breakthrough in performance the Tata system has scored. The reasons for this include the ability of an innovative private enterprise to quickly mobilise resources to exploit the latest off-the-shelf processors and platforms to put together a world class HPC system.

Vijay Bharve, in his blog story has some more info and links. Hindu itself seems to have published the story a couple of days ago which (I, like, totally missed) describes some juicy politics (and in which IISc also makes an appearance):

Interestingly, however, the release from the Tata Group fails to make any mention of the main architect and prime mover of the project, Dr. Narendra Karmarkar, an alumnus of IIT Bombay and formerly of the Tata Institute of Fundamental Research (TIFR), who quit last year following lack of adequate financial support (to the tune of Rs. 400 crore) for his supercomputing project from the institute and joined the Tata Group having found the backing of Ratan Tata and the management of the group (see Business Line, May 6, 2006).

The CRL was established in July 2006 by a group of “like-minded alumni of IIT Bombay,” as the CRL website states.

However, earlier this year, Karmarkar left the Tatas (along with his core team) when they fell apart following differences over the overall plan and the set of goals and objectives of the HPC project, which included that the first system should be given to the Indian government.

But there can be little doubt that the basic idea of the architecture that has been used in the system that has been included in Top 10 belongs to Dr. Karmarkar.

According to Dr. Karmarkar, only five percent of his ideas, which he had shared with the Tatas, have been made use of the current architecture. “With even this, if it can make it to the fourth place, the system can surely make it to No.1 when all my ideas are incorporated,” he says.

“The design and concept took nearly four years, much of it was developed during my days at the TIFR. I have a long and complex plan of how to go about putting all my ideas into an HPC system. But I also have a balanced set of objectives. I am looking for a backer who would not only provide the money but should share the vision and the entire objective,” he added.

The other Indian systems include an IBM eServer Blue Gene Solution of the Indian Institute of Science (IISc), with 18.7 teraflops sustained performance and 23 teraflops peak performance and ranked 58th, another HP Cluster of Tata-CRL with 1440 processors with 9.3 teraflops sustained performance and 15.4 teraflops peak performance and ranked 179th , and six IBM systems in various industrial enterprises, ranked 152nd, 158th, 336th, 339th, 340th and 371st respectively.

I find it curious that the Hindu editorial which compares CDAC’s performance with that of Tata group fails to mention what Kamarkar himself has to say, namely, that

The design and concept took nearly four years, much of it was developed during my days at the TIFR.

Take a look!

### Surfer dudes and stunned physicists

November 15, 2007

When I saw the title of the story (Surfer dude stuns physicists with theory of everything), I thought the link is to that finest news source of America, Onion. But, no! The report seems to be genuine, though, the language does remind you of Onion — like this paragraph for example:

Although the work of 39 year old Garrett Lisi still has a way to go to convince the establishment, let alone match the achievements of Albert Einstein, the two do have one thing in common: Einstein also began his great adventure in theoretical physics while outside the mainstream scientific establishment, working as a patent officer, though failed to achieve the Holy Grail, an overarching explanation to unite all the particles and forces of the cosmos.

Update: A couple of bloggers give their comments on Lisi’s paper and the surrounding hoopla: here is Sean at Cosmic Variance as to why he would rather be reading something else:

The paper seems to involve a novel mix-up between internal symmetries and spacetime symmetries, including adding particles of different spin. This runs against the spirit, if not precisely the letter, of the Coleman-Mandula theorem. Okay, maybe there is a miraculous new way of using loopholes in that theorem to do fun things. But I would be much more likely to invest time trying to understand a paper that was devoted to how we can use such loopholes to mix up bosons and fermions in an unexpected way, and explained clearly why this was possible even though you might initially be skeptical, than in a paper that purports to be a theory of everything and mixes up bosons and fermions so casually.

Rahul at E’s flat, Ah’s flat too wonders if Lisi is a crackpot (and links to several posts for and against the view point):

When such “manifestly crackpot” work can cause such excitement, one is inclined to sympathise with Motl when he frets: ‘Would cranks with their “theories of everything” who know less than 1% what I do and whose IQ is 45 below mine – literally an inferior species – would be placed upon us or even dictate what we can think about physics? Well, this epoch just here…’

Motl observes that the author, Garrett Lisi, is so ignorant of basic physics as to add fermions to bosons, or Grassman numbers to ordinary numbers. As he says, high school students know not to add quantities of different dimensions. So this Lisi guy must be quite a crank.

But the media quotes some well-known physicists — Lee Smolin, for one — as being quite excited by Lisi’s work. And Abhay Ashtekar is quoted here as being receptive to the work, and unconcerned about its defects: “You have to solve problems one at a time.” We’re surrounded by crackpots.

Or maybe Motl missed something? Lisi thinks so.

Take a look!

### HowTo: make grant review process open

November 15, 2007

Better ways have been developed to convey complex ideas. They are called “presentation” and “conversation”.

I want to do what I would do in a business setting. I want to look you in the eye and explain to you why you would be foolish not to fund my proposal; i.e.;

1) that you have a problem,
2) that I know how to solve it
3) that my team has or can find the right people to solve it
4) that those objections which make any sense are already accounted for in the plan

If I can’t look you in the eye, could we at least try instant messaging?

Why do you insist on presentation mechanisms that are practically guaranteed to fail to communicate the ideas and address the objections?

Interesting; take a look!

### Quasicrystals, strings, jamming and solidification

November 15, 2007

Yesterday, I heard Prof. Sharon C Glotzer on From quasicrystals to strings: Unusual ordering processes in jamming and solidification. The talk was split into two parts. The first tried to answer the question of the mechanism behind quasicrystal growth — the short answer: the quasicrystal nuclei eat up the surrounding icosahedral clusters of atoms and grow. The second part tried to show that the microscopic mechanisms behind the glass formation in supercooled liquids and jamming transition in granular systems are probably very similar–the formation of fast moving strings. A short introduction to the dynamical heterogeneity driven structural transitions can be found in Prof. Glotzer’s research page, while the preprint of a paper titled “How do quasicrystals grow?” can be downloaded from this page. Here is my version of the summary of some of the interesting things/concepts/ideas I learnt in the course of the talk.

#### Quasicrystals

If a liquid is cooled, at its freezing temperature, it becomes a solid. However, it is possible to cool a liquid below its freezing temperature and keep it in the liquid state. Such a liquid is called supercooled liquid. There are two processes that can occur in a supercooled liquid: it can either solidify into an ordered, crystalline solid, or, it can freeze into an amorphous glassy phase; and, the transformation from the supercooled liquid to the ordered or glassy phase is a highly path dependent process; many considerations play a role in determining the final outcome, and, packing is one of the most crucial.

##### Packing

Consider spheres which are interacting using a spherically symmetric forces. In 2D, such an interaction leads to hexagonal packing, both globally and locally. On the other hand, in 3D, locally, the atoms pack into icosahedral clusters; however, since the icosahedra cannot fill space, the face centered cubic (fcc) and hexagonal close packed (hcp) phases are usually favoured at the global level. In fact, this competition between the different preferences between long range and short range ordered is what leads to frustration in supercooled liquids eventually leading to glass and quasicrystal formation.

##### Growth of quasicrystals

When a liquid is cooled below its freezing temperature, at a certain temperature the relaxation times become greater than 100 seconds. The temperature at which this change happens is known as the glass transition temperature; in this sense, glass are but slow liquids.

On the other hand, the mechanism behind the formation of an ordered crystalline solid is nucleation and growth. Nucelation is an activated event, and is usually modelled as a chemical reaction, and the nucelus size is one of the most important factors; also, many a times a metastable, intermediate solid nucleates from the supercooled liquid.

The quasicrystalline nucleus is one such metastable phase that nucleates in a supercooled liquid. However, unlike the case of ordered crystal, where the nuclei acts as a template for further growth, it is not clear how local interactions between the quasicrystalline nucleus and supercooled liquid can lead to quasiperiodicity. The answer to the puzzle is that interpenetrating, face sharing icosahedra are formed; these have low energy as well as low mobility. And, these icosahedra play a key role in quasicrystalline growth since the nuclei capture these and turn them into a part of the growing quasicrytal.

##### An aside: Quasicrystalline order parameter

Order parameters and field variables which are typically used in modelling to distinguish between different phases; for example, in solidification studies, the order parameters are rigged in such a way that they take a value of unity (or zero) in the solid and a value of zero (or unity) in the liquid. One such order parameter used to denote fcc, hcp, bcc etc are order parameters which are cluster “shape descriptors” called $q_{6}(i) \cdot q_{6} (j)$. During their study of growth of quasicrystals from supercooled liquid, Glotzer and her co-workers needed an order parameter to distinguish quasicrystals from the surrounding liquid and other solid phases. They have found that similar shape descriptors which are extended to higher order nearest neighbours (2.5) and higher order spherical harmonics (12 for dodecahedral, and 10 for decagonal) serve the purpose admirably for quasicrystals.