Recently there have been a slew of papers on comparison of different schemes to solve Euler equations, especially in the cosmological context. The most recent work has been http://arxiv.org/abs/0901.4107 and other papers by the same group. They claim that an unstructured (and regularized at every timestep) moving mesh combines the best features of SPH (constant in mass resolution, Galilean invariance) and Eulerian codes (well-resolved and contact discontinuity). Moving mesh has an additional advantage: the contact discontinuity is very sharply resolved because of small diffusion error ($\propto v \Delta x$, where $v$ is the fluid velocity in a non-moving mesh). Therefore a moving mesh mode does very well compared to a fixed-grid method for problems with large advection velocities. The SPH codes aren't even capable of qualitatively simulate turbulence; they don't get the inertial range because of the lack of interaction of equal-sized eddies. The reason is that all eddies except the larges ones are diffused by smoothing noise. So SPH is not good to simulate subsonic flows and there is no easy fix for this.
Friday, September 23, 2011
Wednesday, August 31, 2011
Undergraduate degree in Science: IITs vs. IISc?
IISc has started its BS (Bachelors in Science) program this year (http://www.iisc.ernet.in/ug/). I thinks its a great beginning and I hope that the program lives up to its expectations. Comparison of this program with the already mature undergraduate programs at IITs is natural. One feature that makes the IISc program much superior over the IITs is the flexibility to choose majors after first two years. In IITs one has to choose his specialization at the very beginning, based only on the rank and rarely on taste/aptitude. The second advantage that IISc has over IITs, especially if the student wants to go into research in future, is that it was a graduate school till recently, which means that the focus has been mainly on research. Unlike IISc, faculty members have to spend a lot of time in teaching at IITs. The IITs have one big advantage for students interesting in finding jobs right after graduation. They have an extensive alumni network, good reputation, and a well-set placement routine. So the choice is up to you the student; just don't do something because everyone else is doing it!
Thursday, July 07, 2011
Setting up your your mac on IISc's network
Ethernet:
Get an IP address, subnet mask, router, DNS server info from the department system admin. configure IPv4 manually and enter these numbers in the network settings. You'll need to set proxies to connect to the web. Set all (http, https, etc.) proxies; one of the proxy servers is proxy.iisc.ernet.in : 3128 (port number). Get username and password from SERC. you can use a guest username password from the local system admin for the time being. You'll need to fill/submit forms at SERC to get the username/password for the proxy server.
Airport:
Use DHCP to configure IP. 802.1 WPA Enterprise authentication (802.1X is one of the tabs on the panel which appears after you click Advanced tab on the network preferences); create a user profile using the username and password obtained from SERC. You'll need to fill another form for this. There is a wireless network at the main guest house (ghwan) that one can use; I think you'll need some kind of guest username/password for 802.1 authentication (i'm not so sure about this). You probably don't require to connect to the proxy server at the guest house (again i'm not so sure). There are instructions (outdated; and not for a mac) provided at the guest house that you can try to decipher.
Forms to fill at SERC:
There are forms at http://nitss.serc.iisc.ernet.in/ that you can fill to activate VPN to login to IISc machines, activate wireless and ethernet access through your laptop, and to access machines outside the IISc network.
Basically, the upshot is that there are many inconveniences in the name of security. All this could be made much more streamlined and secure!
Installing Matlab from SERC on a Mac:
Goto: http://www.serc.iisc.ernet.in/ComputingFacilities/software/matlab-7.7/unix/brkno4b-1.html.
The instructions for mounting the installer on a Mac given there don't work. Instead try
mount -t smbfs //10.16.2.121/matlab-7.7 /matlab-7.7
after creating the directory matlab-7.7 (don't use -o guest option). You have to use sudo command at every step. Then you just follow the instructions given on their webpage.
Get an IP address, subnet mask, router, DNS server info from the department system admin. configure IPv4 manually and enter these numbers in the network settings. You'll need to set proxies to connect to the web. Set all (http, https, etc.) proxies; one of the proxy servers is proxy.iisc.ernet.in : 3128 (port number). Get username and password from SERC. you can use a guest username password from the local system admin for the time being. You'll need to fill/submit forms at SERC to get the username/password for the proxy server.
Airport:
Use DHCP to configure IP. 802.1 WPA Enterprise authentication (802.1X is one of the tabs on the panel which appears after you click Advanced tab on the network preferences); create a user profile using the username and password obtained from SERC. You'll need to fill another form for this. There is a wireless network at the main guest house (ghwan) that one can use; I think you'll need some kind of guest username/password for 802.1 authentication (i'm not so sure about this). You probably don't require to connect to the proxy server at the guest house (again i'm not so sure). There are instructions (outdated; and not for a mac) provided at the guest house that you can try to decipher.
Forms to fill at SERC:
There are forms at http://nitss.serc.iisc.ernet.in/ that you can fill to activate VPN to login to IISc machines, activate wireless and ethernet access through your laptop, and to access machines outside the IISc network.
Basically, the upshot is that there are many inconveniences in the name of security. All this could be made much more streamlined and secure!
Installing Matlab from SERC on a Mac:
Goto: http://www.serc.iisc.ernet.in/ComputingFacilities/software/matlab-7.7/unix/brkno4b-1.html.
The instructions for mounting the installer on a Mac given there don't work. Instead try
mount -t smbfs //10.16.2.121/matlab-7.7 /matlab-7.7
after creating the directory matlab-7.7 (don't use -o guest option). You have to use sudo command at every step. Then you just follow the instructions given on their webpage.
Thursday, June 30, 2011
Steps to joining IISc as an Assistant Prof.
1) go to health center for medical exam. (ht+wt+urine test+vision test+hernia/hydrocele test) get a signed/stamped letter with approval from the chief medical officer. take this letter to the dept.; this is how you officially join.
2) the dept. will send the paperwork to the main admin. office and you will get further forms to fill out in few days.
3) get in touch with the Estate office (Kannan + Sridhar) for housing assignment. Asst. profs. qualify for D-type (2 BR) quarters. you have to pester them continuously to get the apartment repaired to make it livable. they do not function coherently and you'll have to talk to them very clearly and pester them to get the apartment ready on time.
4) open bank account at campus SBI. get a form from them and fill it out. you need to get a referral from one of your colleagues. you'll need an address proof; department can write a letter saying that you have joined the department and you can use the department address. you'll need two passport-size photos. you'll get a debit/ATM card and a pass-book. don't forget to apply for internet banking.
Its a good idea to get 10 copies of your driver's license and 10 passport-size photos. You need them for everything!
other things:
1) bsnl and airtel have best signal in and around campus. there is a bsnl vendor on campus near the estate office. i find that i am dropping a lot of calls on my prepaid bsnl sim and i am thinking of switching to airtel. i'll update on how airtel does.
2) there are several eating options, both on-campus and outside. i like prakruti veg. canteen near the main gate and the tata book house. tea board is right next to it. then there is faculty club but you'll need your ID to be able to use it; you an ask a colleague to help with lunch. the main guest house serves breakfast, lunch, and dinner but I think you have to be staying in the guest house to be able to use this. on yeshwantpur side there is a fast food place near the yeshwantpur circle/fly-over and rajini andhra restaurant just opposite it. there may be many more options; i'll expand on this as i discover more.
3) you can buy a bicycle at RR bikes near yeshwantpur circle/fly-over; its on SC road (see the map below for subedarchatram Rd.). there are a lot of used-bike dealers on triveni road close to the yeshwantpur fly-over (see map: http://maps.google.com/maps?ll=13.01897,77.557844&spn=0.001672,0.002221&z=19&lci=com.panoramio.all,com.google.webcams). You can also buy a helmet on Triveni Rd. There are a lot of bike repair shops on the same road. Remember prices are negotiable.
i'll add more stuff as i discover more things.
2) the dept. will send the paperwork to the main admin. office and you will get further forms to fill out in few days.
3) get in touch with the Estate office (Kannan + Sridhar) for housing assignment. Asst. profs. qualify for D-type (2 BR) quarters. you have to pester them continuously to get the apartment repaired to make it livable. they do not function coherently and you'll have to talk to them very clearly and pester them to get the apartment ready on time.
4) open bank account at campus SBI. get a form from them and fill it out. you need to get a referral from one of your colleagues. you'll need an address proof; department can write a letter saying that you have joined the department and you can use the department address. you'll need two passport-size photos. you'll get a debit/ATM card and a pass-book. don't forget to apply for internet banking.
Its a good idea to get 10 copies of your driver's license and 10 passport-size photos. You need them for everything!
other things:
1) bsnl and airtel have best signal in and around campus. there is a bsnl vendor on campus near the estate office. i find that i am dropping a lot of calls on my prepaid bsnl sim and i am thinking of switching to airtel. i'll update on how airtel does.
2) there are several eating options, both on-campus and outside. i like prakruti veg. canteen near the main gate and the tata book house. tea board is right next to it. then there is faculty club but you'll need your ID to be able to use it; you an ask a colleague to help with lunch. the main guest house serves breakfast, lunch, and dinner but I think you have to be staying in the guest house to be able to use this. on yeshwantpur side there is a fast food place near the yeshwantpur circle/fly-over and rajini andhra restaurant just opposite it. there may be many more options; i'll expand on this as i discover more.
3) you can buy a bicycle at RR bikes near yeshwantpur circle/fly-over; its on SC road (see the map below for subedarchatram Rd.). there are a lot of used-bike dealers on triveni road close to the yeshwantpur fly-over (see map: http://maps.google.com/maps?ll=13.01897,77.557844&spn=0.001672,0.002221&z=19&lci=com.panoramio.all,com.google.webcams). You can also buy a helmet on Triveni Rd. There are a lot of bike repair shops on the same road. Remember prices are negotiable.
i'll add more stuff as i discover more things.
Monday, June 20, 2011
The Cosmic Baryon Budget
Among many surprising facts of astronomy is that stars contain a negligible fraction of the total mass in the universe. Most mass (~80%) is in the form of the mysterious dark matter. The minority baryons (comprising rest ~20% of the mass) are distributed in different forms (http://arxiv.org/abs/astro-ph/9712020). At the current redshift most baryonic mass is in the X-ray halos of groups and clusters of galaxies; followed by mass in stars (most of the stars are in spheriodal components of galaxies); and by neutral and molecular gas in galaxies. Groups, the biggest baryon reservoirs, are diffuse and extremely difficult to observe in X-rays and therefore their baryon content is quite uncertain.
At high redshift (z~3) the conditions are rather different. The cosmic density is much higher and thus the cooling time is shorter. Moreover, almost all halos are less massive than 1.e11 solar masses, and hence are cooling efficiently. At that redshift most of the baryons are in the form of neutral Hydrogen clouds embedded within ~ 100s of kpc from galaxies. These `clouds' show up in Ly-alpha absorption of background quasars. Some of these clouds are optically thick and are known as damped Ly-alpha absorber (DLAs). The total mass estimated in these clouds is roughly equal to the total baryonic mass at current redshifts. Comparing the baryonic distribution at z=3 and z=0 then suggests that most Ly-alpha absorbing clouds (and stellar outflows generated by stars formed due to the cooling of these filaments) are shock-heated to X-ray temperatures at z=0 and fill the groups and clusters of galaxies.
Moreover, the total baryonic mass fraction at z=0 and z=3 agrees with the baryonic mass fraction calculated from big-bang nucleosynthesis. This implies that most baryonic mass in the universe is in detectable form; this is a good news for us astrophysicists as we can try to account for all of it.
Sunday, June 12, 2011
A week to go!
Wow! The time has come to say good-bye. I am moving to India in a week. There was no time to feel anything till now, but I'll probably feel nostalgic once I am in India. I have spent last 10 happy and productive years in the US but I am very excited about my move to IISc Bangalore. Its the next phase of my life and there is lots to do and achieve.
Tuesday, May 31, 2011
Virial Shocks in Galactic Halos
This paper (Birnboim & Dekel 2004; http://arxiv.org/abs/astro-ph/0302161) performs a global 1-D linear stability analysis of the virial shock in a self-gravitating dark matter halo. The authors relate the pressure and density of the post-shock gas (using $d\ln P/d\ln \rho = \gamma_{\rm eff}$) and perform a stability analysis on the perturbations of the pressure and gravitational forces, assuming hydrostatic equilibrium. $\gamma_{\rm eff}$ depends on the cooling rate; for an adiabatic ideal gas blob $\gamma_{\rm eff}=5/3$. They derive a critical $\gamma_{\rm eff}$ for the shock stability in presence of cooling; only for $\gamma_{\rm eff}>\gamma_{\rm crit}=1.43$ is the post-shock plasma stable. In presence of cooling, the effective adiabatic index is given by
$$\gamma_{\rm eff}=\gamma + \frac{r_s}{2 u_1} \frac{q}{e},$$
where $r_s$ is the shock radius, $u_1$ is the post-shock velocity ($<0$) and $q$ is the cooling rate and $e$ is the internal energy. This instability criterion can also be expressed in terms of the upstream variables and is roughly equivalent to $\frac{t_{\rm ff}}{t_{\rm cool}} \lesssim {\cal O}(1)$. A similar criterion was proposed by Rees & Ostriker 1977 (http://adsabs.harvard.edu/abs/1977MNRAS.179..541R) for efficient cooling of gas to form galaxies. They argued that if $t_{\rm cool}<t_{\rm ff} \sim t_{\rm snd}$, the gas cools to sub-virial temperatures ($10^4$ K), and the cool gas falls on a free-fall time, unimpeded by gas pressure.
Both these papers find that stable/quasi-steady, pressure supported IGM is formed only for a halo mass $\gtrsim 10^{12} M_\odot$. While RO77 and other subsequent papers envisioned a hot, shocked IGM, BD04 showed that a stable virial shock is formed only for $\gtrsim 10^{12} M_\odot$ halos, and that the gas accretes in the code mode (at $\sim 10^4$ K) for lower mass halos, as observed in recent numerical simulations. BD04 also performed 1-D Lagrangian simulations of shock formation with DM + gas, starting from cosmologically consistent initial conditions. Their numerical simulations agree with their linear stability analysis. The critical halo mass for forming stable shock ($\sim 10^{11.5} M_\odot$) is fairly insensitive to metallicity, angular momentum, baryon fraction, power spectrum, etc. An important effect that BD04 may have missed is feedback heating; in presence of feedback heating the virial shock can become stable as feedback heating provides the post-shock pressure to support the shock.
Recently Dekel and collaborators have argued that cold streams are responsible for star-forming galaxies (SFGs) with SFR~100$M_\odot {\rm yr}^{-1}$; however major mergers are likely responsible for the rarer sub-mm galaxies at z~2 with SFR~1000$M_\odot {\rm yr}^{-1}$. Major mergers are responsible for $< 1/3$ in total star formation.
$$\gamma_{\rm eff}=\gamma + \frac{r_s}{2 u_1} \frac{q}{e},$$
where $r_s$ is the shock radius, $u_1$ is the post-shock velocity ($<0$) and $q$ is the cooling rate and $e$ is the internal energy. This instability criterion can also be expressed in terms of the upstream variables and is roughly equivalent to $\frac{t_{\rm ff}}{t_{\rm cool}} \lesssim {\cal O}(1)$. A similar criterion was proposed by Rees & Ostriker 1977 (http://adsabs.harvard.edu/abs/1977MNRAS.179..541R) for efficient cooling of gas to form galaxies. They argued that if $t_{\rm cool}<t_{\rm ff} \sim t_{\rm snd}$, the gas cools to sub-virial temperatures ($10^4$ K), and the cool gas falls on a free-fall time, unimpeded by gas pressure.
Both these papers find that stable/quasi-steady, pressure supported IGM is formed only for a halo mass $\gtrsim 10^{12} M_\odot$. While RO77 and other subsequent papers envisioned a hot, shocked IGM, BD04 showed that a stable virial shock is formed only for $\gtrsim 10^{12} M_\odot$ halos, and that the gas accretes in the code mode (at $\sim 10^4$ K) for lower mass halos, as observed in recent numerical simulations. BD04 also performed 1-D Lagrangian simulations of shock formation with DM + gas, starting from cosmologically consistent initial conditions. Their numerical simulations agree with their linear stability analysis. The critical halo mass for forming stable shock ($\sim 10^{11.5} M_\odot$) is fairly insensitive to metallicity, angular momentum, baryon fraction, power spectrum, etc. An important effect that BD04 may have missed is feedback heating; in presence of feedback heating the virial shock can become stable as feedback heating provides the post-shock pressure to support the shock.
Recently Dekel and collaborators have argued that cold streams are responsible for star-forming galaxies (SFGs) with SFR~100$M_\odot {\rm yr}^{-1}$; however major mergers are likely responsible for the rarer sub-mm galaxies at z~2 with SFR~1000$M_\odot {\rm yr}^{-1}$. Major mergers are responsible for $< 1/3$ in total star formation.
Subscribe to:
Posts (Atom)