Wednesday, December 14, 2011
Mathematica
Attended a one-day training in Long Hanborough on the use of Mathematica by Wolfram expert Jon McLoone. You can certainly do some cool stuff with Mathematica if you put aside the issues of "open source" and "free software" (free as in free speech but it's also pretty expensive to get the software!)
Tuesday, December 13, 2011
First Snow
Snow's falling though quite not enough so it's all melting by the time it hits the ground. But it's snow!
Saturday, November 19, 2011
Thursday, November 10, 2011
Physics Pieces (Part 5)
Interesting set of seminars last week.
Claudio Castelnovo from Royal Holloway, who was a post-doc in Oxford, gave a talk about magnetic monopoles in spin ice, which was featured on the cover of Nature in 2008. He gave a brief theoretical introduction and described results by Morris et al. from neutron scattering experiments with an applied magnetic field to compare against this model of emergent electrodynamics in a 3D condensed matter system.
Daniel Khomskii from Koeln University, Germany spoke on frustrated Mott insulators. The virtual hopping of Hubbard spins around a triangle in a triangular lattice gives an orbital current in third order perturbation theory, which was a somewhat analogous result to magnetic monopoles in spin ice (electric charge instead of magnetic monopoles).
Dieter Jaksch, associated with the atomic and laser physics sub-department of Oxford, and with the Centre for Quantum Technologies in Singapore, gave a talk about strongly driven quantum systems. He described the application of tensor network and path sum approaches to their areas of research, on impurities in optical lattices, Rydberg lattices and tetrahertz pump-probes.
Martin Rees gave another talk, this time on real and counterfactual universes. This was the 8th Dennis Sciama memorial lecture, and Rees was Sciama's graduate student. Stephen Hawking had been Sciama's graduate student for about 2 years when Rees joined the research group. Interestingly, Roger Penrose, who made contributions to astrophysics during the 1960s, was at the talk as well.
Rees gave some of the history of the steady-state theory versus big bang theory in the 1940s and 50s, and spoke about the "tuning" problem of the cosmological constant and the amplitude of fluctuations during the big bang. To produce a universe like ours, there seem to be strong constraints on the allowed value of the cosmological constant (cosmic expansion rate), and some sensitivity also to the fluctuations - too much fluctuations and there are too many black holes, too little and there are too few stars. And gravity needs to be weak so the universe has enough time to become interesting.
Claudio Castelnovo from Royal Holloway, who was a post-doc in Oxford, gave a talk about magnetic monopoles in spin ice, which was featured on the cover of Nature in 2008. He gave a brief theoretical introduction and described results by Morris et al. from neutron scattering experiments with an applied magnetic field to compare against this model of emergent electrodynamics in a 3D condensed matter system.
Daniel Khomskii from Koeln University, Germany spoke on frustrated Mott insulators. The virtual hopping of Hubbard spins around a triangle in a triangular lattice gives an orbital current in third order perturbation theory, which was a somewhat analogous result to magnetic monopoles in spin ice (electric charge instead of magnetic monopoles).
Dieter Jaksch, associated with the atomic and laser physics sub-department of Oxford, and with the Centre for Quantum Technologies in Singapore, gave a talk about strongly driven quantum systems. He described the application of tensor network and path sum approaches to their areas of research, on impurities in optical lattices, Rydberg lattices and tetrahertz pump-probes.
Martin Rees gave another talk, this time on real and counterfactual universes. This was the 8th Dennis Sciama memorial lecture, and Rees was Sciama's graduate student. Stephen Hawking had been Sciama's graduate student for about 2 years when Rees joined the research group. Interestingly, Roger Penrose, who made contributions to astrophysics during the 1960s, was at the talk as well.
Rees gave some of the history of the steady-state theory versus big bang theory in the 1940s and 50s, and spoke about the "tuning" problem of the cosmological constant and the amplitude of fluctuations during the big bang. To produce a universe like ours, there seem to be strong constraints on the allowed value of the cosmological constant (cosmic expansion rate), and some sensitivity also to the fluctuations - too much fluctuations and there are too many black holes, too little and there are too few stars. And gravity needs to be weak so the universe has enough time to become interesting.
Physics Pieces (Part 4)
Talks by two Princeton professors who visited Oxford last week. Dan Marlow gave a talk about luminosity measurements at the Large Hadron Collider, discussing the various technical issues involved when pushing equipment to its limits to test the limits of science. Duncan Haldane spoke about geometry and topological phases of matter, starting from the Gauss-Bonnet theorem and the generalisation by Chern. Interesting history about how Karplus and Luttinger's paper in 1954 (Phys. Rev. 95, 1154–1160) on the anomalous Hall effect was discredited and finally vindicated in 2005.
Fellow graduate student Curt von Keyserlingk gave a talk on the Toric Code too, a toy model involving plaquette and vertex terms in the hamiltonian for the spins. There's a nice pictorial representation of the states of the system, and similar models can lead to what are known as "anyon statistics" for quasiparticles that are neither fermions nor bosons.
Roser Valenti from Goethe University in Frankfurt gave a talk on frustration, using density functional theory to come up with effective models and studying them with quantum Monte Carlo and other methods. Interesting attempt to understand caesium copper chloride and caesium copper bromide by substituting the chlorine atoms with bromine atoms and vice versa.
Martin Rees from Cambridge delivered a public lecture on "The Limits of Science" last week too, and I took the kids to it. He had some encouragement for beginning scientists, who might feel that so many major advances have been made that it's hard to push further forward. His idea was that each advance leads to new questions, and the computational power we have at our disposal is a powerful tool that previous generations did not have. Yet there are likely to be things beyond our ability to comprehend, which require posthuman intelligence. There is a huge gulf between what we can do today and what we should do, whether in research or in public policy, and we need to choose wisely with a longer term view of leaving a fair inheritance for future generations.
Fellow graduate student Curt von Keyserlingk gave a talk on the Toric Code too, a toy model involving plaquette and vertex terms in the hamiltonian for the spins. There's a nice pictorial representation of the states of the system, and similar models can lead to what are known as "anyon statistics" for quasiparticles that are neither fermions nor bosons.
Roser Valenti from Goethe University in Frankfurt gave a talk on frustration, using density functional theory to come up with effective models and studying them with quantum Monte Carlo and other methods. Interesting attempt to understand caesium copper chloride and caesium copper bromide by substituting the chlorine atoms with bromine atoms and vice versa.
Martin Rees from Cambridge delivered a public lecture on "The Limits of Science" last week too, and I took the kids to it. He had some encouragement for beginning scientists, who might feel that so many major advances have been made that it's hard to push further forward. His idea was that each advance leads to new questions, and the computational power we have at our disposal is a powerful tool that previous generations did not have. Yet there are likely to be things beyond our ability to comprehend, which require posthuman intelligence. There is a huge gulf between what we can do today and what we should do, whether in research or in public policy, and we need to choose wisely with a longer term view of leaving a fair inheritance for future generations.
Tuesday, November 8, 2011
gNewSense
Sunday, November 6, 2011
rms
Richard Stallman came by to Oxford to give a talk ("A Free Digital Society") on Friday. As I was at another seminar, I didn't manage to attend even though it was relatively conveniently held at the physics department. As a result, I read up more about his life and work, after finding this very interesting piece that details his preferences and requests for anyone interested in getting him to speak.
The actual talk he gave was probably more or less along the lines of this earlier transcript. I agree that indeed, copyright and patents are relics of the past that now mainly benefit corporations only and stand in the way of progress. We do need to disentangle the issues to both encourage sharing and cooperation and also encourage content creation.
Reading up again about his life and cause reminded me of the ideals of freedom, choice and software. In many ways, universities embody the essence of freedom and cooperation - research published in open journals are analogous to "copyleft".
I'll end this post by pointing out that the story of his younger days is pretty interesting. He did physics at Harvard and spent weekends at the MIT Artificial Intelligence lab, where the "freedom" issue first reared its head in the form of a proprietary printer driver. He was a physics graduate student for a while and decided that he could make a greater contribution through computer code.
[On a slightly more technical/purist point, as I'm using Ubuntu GNU/Linux now with non-free drivers, I've sacrificed some "freedom" for the convenience of getting my laptop to work well. Just downloaded the fully free gNewSense which I'll test-drive.]
The actual talk he gave was probably more or less along the lines of this earlier transcript. I agree that indeed, copyright and patents are relics of the past that now mainly benefit corporations only and stand in the way of progress. We do need to disentangle the issues to both encourage sharing and cooperation and also encourage content creation.
Reading up again about his life and cause reminded me of the ideals of freedom, choice and software. In many ways, universities embody the essence of freedom and cooperation - research published in open journals are analogous to "copyleft".
I'll end this post by pointing out that the story of his younger days is pretty interesting. He did physics at Harvard and spent weekends at the MIT Artificial Intelligence lab, where the "freedom" issue first reared its head in the form of a proprietary printer driver. He was a physics graduate student for a while and decided that he could make a greater contribution through computer code.
[On a slightly more technical/purist point, as I'm using Ubuntu GNU/Linux now with non-free drivers, I've sacrificed some "freedom" for the convenience of getting my laptop to work well. Just downloaded the fully free gNewSense which I'll test-drive.]
Tuesday, November 1, 2011
London
Visited London over the weekend with the family (photos). We went to the Natural History Museum. Plenty of dinosaur bones and other pieces from our past. Didn't have time for the Science Museum - maybe next time!
Physics Pieces (Part 3)
It's a new month!
This is going to be a short piece since I've already reported on some of the talks from last week. There was a talk by Jacques Laskar from IMCCE in Paris. Unfortunately I missed this one but there is a video that he showed at the end of the talk.
Joanna Haigh from Imperial College London gave a talk about how changes in the sun affect climate. The prevalence of sunspots has a 11-year cycle of fluctuations, and recent data has suggested a spectral distribution quite different from what is usually assumed.
This is going to be a short piece since I've already reported on some of the talks from last week. There was a talk by Jacques Laskar from IMCCE in Paris. Unfortunately I missed this one but there is a video that he showed at the end of the talk.
Joanna Haigh from Imperial College London gave a talk about how changes in the sun affect climate. The prevalence of sunspots has a 11-year cycle of fluctuations, and recent data has suggested a spectral distribution quite different from what is usually assumed.
Wednesday, October 26, 2011
Physics Pieces (Part 2)
Another exciting week has gone by, and the current week is unrelentless in seeking to outdo the last.
It's just mid-week and already David Weiss from Penn State gave us a summary of the various cold atom experiments going on in his lab - ranging from quantum computing on a 3-d lattice to a quantum version of "Newton's cradle" and a search for the electron dipole moment.
Fiona Burnell, a post-doc in our group, gave an informal presentation on a toy model for fractional topological insulators. Useful ideas to help me navigate the zoo of condensed matter systems out there. She did her PhD at Princeton under Shivaji Sondhi and had started her studies there while I was still there as an undergraduate.
But the real focus of this post was last week.
Alessandro de Silva from ICTP Trieste spoke about quantum quenches, studying the behaviour of systems quickly thrown from an equilibrium state to another via a non-equilibrium process.
Chris Hooley from St Andrews gave an entertaining lecture, based on some initial research into the use of complex temperatures for the thermodynamic partition function. It was useful for me as he spent a lot of time motivating the discussion and talking about quantum critical points.
David Nelson from Harvard gave a very accessible presentation on the packing problem on a curved surface. Very nice simulations and visuals for a practical and mathematical topic, and actually had everyone in the audience looking at a golf ball (the distribution of the dimples on the surface is related to how viruses arrange their capsomeres).
Lev Ioffe from Rutgers talked about quantum coherence, and what I really enjoyed were the introductory graphics with spherical cows and the philosophical idea that if the environment itself is quantum, then how does quantum decoherence set in? So it's like Schrödinger's cats looking at Schrödinger's cats.
With the complement of courses and problem sets, I am slowly but surely progressing on the path of understanding nature better!
It's just mid-week and already David Weiss from Penn State gave us a summary of the various cold atom experiments going on in his lab - ranging from quantum computing on a 3-d lattice to a quantum version of "Newton's cradle" and a search for the electron dipole moment.
Fiona Burnell, a post-doc in our group, gave an informal presentation on a toy model for fractional topological insulators. Useful ideas to help me navigate the zoo of condensed matter systems out there. She did her PhD at Princeton under Shivaji Sondhi and had started her studies there while I was still there as an undergraduate.
But the real focus of this post was last week.
Alessandro de Silva from ICTP Trieste spoke about quantum quenches, studying the behaviour of systems quickly thrown from an equilibrium state to another via a non-equilibrium process.
Chris Hooley from St Andrews gave an entertaining lecture, based on some initial research into the use of complex temperatures for the thermodynamic partition function. It was useful for me as he spent a lot of time motivating the discussion and talking about quantum critical points.
David Nelson from Harvard gave a very accessible presentation on the packing problem on a curved surface. Very nice simulations and visuals for a practical and mathematical topic, and actually had everyone in the audience looking at a golf ball (the distribution of the dimples on the surface is related to how viruses arrange their capsomeres).
Lev Ioffe from Rutgers talked about quantum coherence, and what I really enjoyed were the introductory graphics with spherical cows and the philosophical idea that if the environment itself is quantum, then how does quantum decoherence set in? So it's like Schrödinger's cats looking at Schrödinger's cats.
With the complement of courses and problem sets, I am slowly but surely progressing on the path of understanding nature better!
Tuesday, October 25, 2011
Hibernation
It's getting so cold in Oxford. And it's going to get colder.
On hibernation, somehow the upgrade to Ubuntu 11.10 broke hibernation, which was working when I was using Ubuntu 11.04. Or perhaps it was from my fiddling with the fglrx driver for AMD graphics.
Got a message that there wasn't enough space, which was weird. Followed the instructions on making a new swap partition work for hibernate, and even tried to create a 10 GB swap partition, but was still being told that there wasn't enough memory! It only worked intermittently when I had freshly booted in.
Stumbled upon this forum thread which led me to the answer.
On hibernation, somehow the upgrade to Ubuntu 11.10 broke hibernation, which was working when I was using Ubuntu 11.04. Or perhaps it was from my fiddling with the fglrx driver for AMD graphics.
Got a message that there wasn't enough space, which was weird. Followed the instructions on making a new swap partition work for hibernate, and even tried to create a 10 GB swap partition, but was still being told that there wasn't enough memory! It only worked intermittently when I had freshly booted in.
Stumbled upon this forum thread which led me to the answer.
- Checked /var/log/syslog which showed that the hibernation image needed more pages than available in the swap partition.
- The magic fix is "echo 0 > /sys/power/image_size", which was set at too high a value on my system.
- As this value is reset across reboots, I embedded this line at the start of /usr/sbin/pm-hibernate.
Tuesday, October 18, 2011
Physics Pieces (Part 1)
A quick review of the week gone by...
A talk given by Olga Sikora (who worked in Nic Shannon's group in Bristol) was on spin ice and mentioned that frustration can occur as "charge frustration" as well. Professor Shannon is based in Oxford this year, and is subsequently headed to the new Okinawa Institute of Science and Technology next year.
A talk by Shivaji Sondhi presented a very interesting description of quantum liquids by considering the vortex structure induced in superfluids when coupled with real-space rotation in a bucket. Professor Sondhi was my advisor for my very first research project on the Ising model during Junior Year at Princeton.
A talk by Mark Newman on networks. Excellent for showing the application of physical insight (mainly statistical physics) into problems such as social networks, such as a simple proof of why it usually is that your friends on average have more friends than you (Ref). Professor Newman is a prolific author and did his DPhil at Oxford, co-supervised by Robin Stinchcombe and David Sherrington.
A talk given by Olga Sikora (who worked in Nic Shannon's group in Bristol) was on spin ice and mentioned that frustration can occur as "charge frustration" as well. Professor Shannon is based in Oxford this year, and is subsequently headed to the new Okinawa Institute of Science and Technology next year.
A talk by Shivaji Sondhi presented a very interesting description of quantum liquids by considering the vortex structure induced in superfluids when coupled with real-space rotation in a bucket. Professor Sondhi was my advisor for my very first research project on the Ising model during Junior Year at Princeton.
A talk by Mark Newman on networks. Excellent for showing the application of physical insight (mainly statistical physics) into problems such as social networks, such as a simple proof of why it usually is that your friends on average have more friends than you (Ref). Professor Newman is a prolific author and did his DPhil at Oxford, co-supervised by Robin Stinchcombe and David Sherrington.
Monday, October 17, 2011
Thunderstorm
The first real rain since we arrived. Good thing it's nighttime and we're indoors sleeping.
Cycling
First bicycle accident today. The form factor of a stationary car fluctuates by the size of a door when there is at least one occupant.
Sunday, October 16, 2011
Ubuntu 11.10
Just upgraded to Ubuntu 11.10, what I appreciate is good support for laptop hardware. I even get two-finger scroll with the touchpad out of the box.
Matriculation
The process of becoming a member of the university, tied to past and future generations by a ritual of Latin pronouncements.
Tuesday, October 11, 2011
ring flinger
parting the moss
reaching an unadorned hand
into the cold water
turning up
a bottle cap
soiled in the murky depths
a moment ago like a fish
splashing off a fat worm
fingers now in again
caressing the river bed
for the missing ring
sensation running
backwards in time
an anchor sinking into the cherwell
flashing back to the jeweler's
and tossed
upon the dreaming spires
reaching an unadorned hand
into the cold water
turning up
a bottle cap
soiled in the murky depths
a moment ago like a fish
splashing off a fat worm
fingers now in again
caressing the river bed
for the missing ring
sensation running
backwards in time
an anchor sinking into the cherwell
flashing back to the jeweler's
and tossed
upon the dreaming spires
Thursday, May 12, 2011
Debian Live
Debian Live is pretty mature with the squeeze release. I downloaded an image to copy to a usb flash drive. Had to make a few tweaks to get it working the way I wanted. With my first post on this blog, I'm going to attempt a brief walkthrough (assuming you already run GNU/Linux):
Step 1
Download an *.img file from
http://cdimage.debian.org/cdimage/release/current-live/i386/usb-hdd/
(assuming you plan to use a computer with an intel processor)
The LXDE version is the smallest. That version and the XFCE version can fit on a 1GB usb drive. The GNOME and KDE versions require a 2GB or larger drive.
Step 2
Follow the manual below, but there are extra steps I needed.
http://live.debian.net/manual/en/html/the-basics.html
(a) Use "dmesg" or "ls -l /dev/disk/by-id" to identify your usb drive. Suppose your drive is "/dev/sdc". Be paranoid in ensuring you have the right identification or risk wiping out a hard disk later!
(b) What you need to do is then unmount the usb drive (assuming it has existing partitions already), without "safely ejecting" it.
So run "umount /dev/sdc1" and "umount /dev/sdc2", etc. as necessary.
(c) Now you are ready to wipe out the contents of the usb drive and load the Debian Live system onto the drive. This will override all contents on the device you indicate below!
Example: "dd if=debian-live-6.0.1-i386-gnome-desktop.img of=/dev/sdc"
The image must be written to the device, not a partition on the device.
Step 3
The next thing is to use the rest of the space on the usb drive to contain your files, which is known as "persistence". I think I had some trouble the first few times because I immediately tried to create a partition out of the remaining space after Step 2.
So, what you should do is "safely eject" the usb drive first, then plug it back in again. Use "palimpsest" or "gparted" or similar disk utilities to create an ext4 partition out of the remaining space, and give the partition the label "home-rw".
This label name for the partition is special, and Debian Live will know to automagically mount it as the "/home" partition and fold it seamlessly into the filesystem.
Step 4
The final step is to enable persistence as a default. When you plug in the usb drive into a running Linux system (or even a Windows system, since the partition containing the Debian Live image is a FAT partition), go into the "syslinux" folder and edit the text file "live.cfg". Add the word "persistent" (without the quotes) to the end of the last line in the code blocks for the labels "Live" and "Live 686". Save the file and you're done!
Testing
To use or test your Debian Live, plug it in to a computer before switching it on. You'll need a computer that allows you to choose to boot from the usb drive. Once you do that, everything should work when you choose either the "Live" or the "Live 686" option in the graphical Debian boot prompt. In case you're wondering, "686" is for dualcore and more advanced processors. Enjoy!
[posting from Debian Live]
Step 1
Download an *.img file from
http://cdimage.debian.org/cdimage/release/current-live/i386/usb-hdd/
(assuming you plan to use a computer with an intel processor)
The LXDE version is the smallest. That version and the XFCE version can fit on a 1GB usb drive. The GNOME and KDE versions require a 2GB or larger drive.
Step 2
Follow the manual below, but there are extra steps I needed.
http://live.debian.net/manual/en/html/the-basics.html
(a) Use "dmesg" or "ls -l /dev/disk/by-id" to identify your usb drive. Suppose your drive is "/dev/sdc". Be paranoid in ensuring you have the right identification or risk wiping out a hard disk later!
(b) What you need to do is then unmount the usb drive (assuming it has existing partitions already), without "safely ejecting" it.
So run "umount /dev/sdc1" and "umount /dev/sdc2", etc. as necessary.
(c) Now you are ready to wipe out the contents of the usb drive and load the Debian Live system onto the drive. This will override all contents on the device you indicate below!
Example: "dd if=debian-live-6.0.1-i386-gnome-desktop.img of=/dev/sdc"
The image must be written to the device, not a partition on the device.
Step 3
The next thing is to use the rest of the space on the usb drive to contain your files, which is known as "persistence". I think I had some trouble the first few times because I immediately tried to create a partition out of the remaining space after Step 2.
So, what you should do is "safely eject" the usb drive first, then plug it back in again. Use "palimpsest" or "gparted" or similar disk utilities to create an ext4 partition out of the remaining space, and give the partition the label "home-rw".
This label name for the partition is special, and Debian Live will know to automagically mount it as the "/home" partition and fold it seamlessly into the filesystem.
Step 4
The final step is to enable persistence as a default. When you plug in the usb drive into a running Linux system (or even a Windows system, since the partition containing the Debian Live image is a FAT partition), go into the "syslinux" folder and edit the text file "live.cfg". Add the word "persistent" (without the quotes) to the end of the last line in the code blocks for the labels "Live" and "Live 686". Save the file and you're done!
Testing
To use or test your Debian Live, plug it in to a computer before switching it on. You'll need a computer that allows you to choose to boot from the usb drive. Once you do that, everything should work when you choose either the "Live" or the "Live 686" option in the graphical Debian boot prompt. In case you're wondering, "686" is for dualcore and more advanced processors. Enjoy!
[posting from Debian Live]
Subscribe to:
Posts (Atom)