I believe I have discovered the difference between the source terms I was deriving and those presented in Skinner & Ostriker. It is emphasized that the source terms presented in that paper are used only for the reconstruction step. Those source terms are derived from simplified versions of the MHD equations which make use of div B = 0 (see example at bottom of this post).

If you leave the MHD equations in their full flux-conserving form, then you get the source terms that I previously posted. Below, I have summarized all the possible source terms from these two methods. "full" corresponds to using the complete flux version of the MHD equations, and "divergence-free" corresponds to what you get when you first expand the MHD equations and take out div B.

Conservative Variables

From this point on, these source terms are written to conveniently compare with cylindrical.f90 in astrobear. The "actual" source terms are these values multiplied by a factor of -1/r. Also, the magnetic terms usually come with a factor of 1/4pi, but all these values are in computational units.

variable	full source term	divergence-free source term
\rho	\rho v_r	\rho v_r
\rho v_r	\rho(v_{r}^{2} - v_{\phi}^{2}) - B_{r}^{2} + B_{\phi}^{2}	\rho(v_{r}^{2} - v_{\phi}^{2}) + B_{\phi}^{2}
\rho v_\phi	2 \rho v_{r} v_{\phi} - 2 B_{r} B_{\phi}	2 \rho v_{r} v_{\phi} - B_{r} B_{\phi}
\rho v_z	\rho v_{r} v_{z} - B_{r} B_{z}	\rho v_{r} v_{z}
E	v_{r} (E + P + \frac{B^{2}}{2}) - B_{r} (\vec{v} \cdot \vec{B})	v_{r} (E + P + \frac{B^{2}}{2})
B_r	0	v_{r} B_{r}
B_\phi	0	v_{\phi} B_{r}
B_z	v_{r} B_{z} - v_{z} B_{r}	v_{r} B_{z}

Primitive Variables

Momentum —→ velocity which is not as simple as just dividing the momentum source terms by the density. Those have to take into account a product rule that introduces another source term. The density and magnetic field source terms remain the same. Energy —→ thermal pressure which also has to be derived on its own.

variable	full source term	divergence-free source term
\rho	\rho v_r	\rho v_r
v_r	- v_{\phi}^{2} - \frac{B_{r}^{2}}{\rho} + \frac{B_{\phi}^{2}}{\rho}	- v_{\phi}^{2} + \frac{B_{\phi}^{2}}{\rho}
v_\phi	v_{r} v_{\phi} - \frac{2 B_{r} B_{\phi}}{\rho}	v_{r} v_{\phi} - \frac{B_{r} B_{\phi}}{\rho}
v_z	- \frac{B_{r} B_{z}}{\rho}	0
P	\gamma P v_{r} + (\gamma - 1) B_{r} (\vec{v} \cdot \vec{B})	\gamma P v_{r}
B_r	0	v_{r} B_{r}
B_\phi	0	v_{\phi} B_{r}
B_z	v_{r} B_{z} - v_{z} B_{r}	v_{r} B_{z}

Now the only part that is inconsistent with Skinner & Ostriker is the divergence-free version of the Br source term.

MHD Contributions in the Momentum/Velocity Equation

In the momentum equation, there is a term that looks like: \vec{\nabla} \cdot ( \vec{B} \otimes \vec{B}) If you brute force this equation and take the gradient of this tensor, you get the magnetic contributions for the momentum/velocity source terms as shown in the "full" columns above. However, you can also use a tensor identity to rewrite this expression as follows: \vec{B}(\vec{\nabla} \cdot \vec{B}) + (\vec{B} \cdot \vec{\nabla})\vec{B} Then, you could get rid of the div B term since div B should = 0. However, there are source terms associated with div B, so your overall source terms have now changed. Using this method results in the "divergence-free" columns above.

Similarly, the MHD contributions for energy/pressure and B also change if you use other calculus identities to expand the magnetic terms and get rid of any div B terms.

• Runs

Should be done tomorrow.

• Movies:
  

1 Jupiter Mass ~3 orbits (complete) ​movie
5 Jupiter Mass 5 orbits (complete) ​movie
10 Jupiter Mass 10 orbits (almost complete) ​movie
20 Jupiter Mass 10 orbits (almost complete)

• Bug

Appears to be solved with threading disabled.
Created ticked to document the issue ​268

• Getting ready for trip to Berkeley/Caltech
• Christina's setup takes 2 months on 4096 bluestreak cores. Scaling is ok - so likely a combination of
  • higher filling fractions
  • MHD
  • Bluestreak speed issues.

Can't look at mesh since didn't get first frame. Another job with a higher frame rate has been in the queue since Saturday morning.

• Golden Version & Blue Streak
  1. AstroBEAR performance on blue streak: ​https://clover.pas.rochester.edu/trac/astrobear/blog/johannjc11132012#comment-1
  2. Still have problem for modules GravitationalCascade (#265), IsotropicTurbulence (#266) and SlowMolecularCloudFormation

• Users
  1. IO will skype in during the meeting
  2. Will ask Rich for Shaz's local account.

• Start working on MUSCL

Circum Nuclear Disk (CND) in the Galactic Center (GC)

Brief Intro: CND is a Keplerian ring of inner radius 1pc and outer radii 3-7 pc or more (probably not more) around the supermassive blackhole in the GC, having a mass ~4.2e6 M_sun. It contains 10e4 - 10e5 solar masses of very warm, turbulent and clumpy molecular and atomic gas. There is a luminous young cluster of massive stars located within a few tenths of a pc from the blackhole, outputting strong spherical stellar wind. A torodial magnetic field of 1 milliGauss exists in the disk.

Run time and Environment: A batch job was submitted and run in the UCLA *Hoffman2 Cluster. 40 CPUs were requested and It took ~1 week on queuing. Interactive mode with maximum CPU=8 and maximum run time=24hours also available for users. [*The maximum # of CPUs and maximum run time that a normal user can request for are 128 and 24 hours respectively] The total integration time was only 15% of a revolution time on the outer edge of the disk (t~54,000 yr & total frames=20). The computational run time was ~15.5 hours. 40 cores were used.

Setup of the Problem: (Problem was set up with the BasicDisk module by Jonathan) A molecular ring with of an inner radius of 1pc and an outer radius of 4pc, and a constant thickness of 1pc was set up. To prevent inflow/outflow of matter in the z direction to the disk, the disk and the envelope are in hydrostatic and pressure equilibrium . The outflow module was used to simulate the ionized stellar wind from the center of the ring. The ionization module was also used. Moreover, we imposed a torodial B field of magnitude 1mG in the disk. 

Some important  global/physical parameters are shown as the following: Resolution: 32, 32, 16 AMRlevel: 2

Outflow

• outflow radius: 0.4pc (from the central blackhole)
• outflow velocity: 700km/s
• outflow density: *1000 /cc

(*from observation: total mass loss rate from the central cluster ~3e-2 solar masses per year)

• outflow temperature: 1e4 K

Disk

• density: 1e5 /cc

(=total mass ~2.5e4 Msun)

• temperature: 300 K 

Result: 1) Flaring Disk started to form in the 3rd frame (t=8,100 yr) after the ionized wind hit the inner edge of the CND

2) As the outflow compressed the inner region of the disk, dense equatorial region formed

3) A density wave propagating outwards radially, a second wave was seen in the 18th frame.

4) Rotational KH instability? (4-fold symmetrical feature that was attributable to the gridding, need a further check with better resolution)

Image of the simulation attached (Face-on & Edge-on)

I have been double checking the cylindrical source terms in astrobear, because the Athena paper (​link) that I was referencing had some differences. I derived the source terms myself…this time using the full 3D MHD equations, and my results were different from Skinner and Ostriker. Astrobear has two cylindrical source term routines for both the conservative and primitive formulations. I get the same results as the conservative routine. However, the primitive routine has some errors, but those are easy to fix. Below is a table showing what I believe the source terms should be…

Conservative

From this point on, these source terms are written to conveniently compare with cylindrical.f90 in astrobear. The "actual" source terms are these values multiplied by a factor of -1/r. Also, the magnetic terms usually come with a factor of 1/4pi, but all these values are in computational units.

variable	source term
\rho	\rho v_r
\rho v_r	\rho(v_{r}^{2} - v_{\phi}^{2}) - B_{r}^{2} + B_{\phi}^{2}
\rho v_\phi	2 \rho v_{r} v_{\phi} - 2 B_{r} B_{\phi}
\rho v_z	\rho v_{r} v_{z} - B_{r} B_{z}
E	v_{r} (E + P + \frac{B^{2}}{2}) - B_{r} (\vec{v} \cdot \vec{B})
B_r	0
B_\phi	0
B_z	v_{r} B_{z} - v_{z} B_{r}

Primitive

Momentum —→ velocity which is not as simple as just dividing the momentum source terms by the density. Those have to be derived separately (see Jonathan's comment). The density and magnetic field source terms remain the same. Energy —→ thermal pressure which also has to be derived on its own.

variable	source term
\rho	\rho v_r
v_r	- v_{\phi}^{2} - \frac{B_{r}^{2}}{\rho} + \frac{B_{\phi}^{2}}{\rho}
v_\phi	v_{r} v_{\phi} - \frac{2 B_{r} B_{\phi}}{\rho}
v_z	- \frac{B_{r} B_{z}}{\rho}
P	\gamma P v_{r} + (\gamma - 1) B_{r} (\vec{v} \cdot \vec{B})
B_r	0
B_\phi	0
B_z	v_{r} B_{z} - v_{z} B_{r}

So in conclusion, I guess it's possible that I am not understanding some of Skinner and Ostriker's assumptions and/or techniques, but I think they might have some errors in their source terms.

CRL 618

Movies:

Clump: ​http://www.pas.rochester.edu/~martinhe/2012/crl20nov.gif

Working on the triggered star formation project. By changing the ambient density, collapse could be triggered for a super critical BE sphere. The BE sphere has a cut at 0.15 pc, the density at the outer edge is always lower than the BE profile. See the following movies:

20 times denser ambient:
​http://www.pas.rochester.edu/~shuleli/collapse/heavy20_2d.gif
line cut:
​http://www.pas.rochester.edu/~shuleli/collapse/heavy20_line.gif
10 times denser ambient:
​http://www.pas.rochester.edu/~shuleli/collapse/heavy10_2d.gif
For any ambient with ambient density < 8, the sphere simply dissipates:
5 times denser ambient:
​http://www.pas.rochester.edu/~shuleli/collapse/heavy5_2d.gif
normal ambient:
​http://www.pas.rochester.edu/~shuleli/collapse/normal.gif
normal ambient line cut:
​http://www.pas.rochester.edu/~shuleli/collapse/normal_line.gif
ambient 5 times lighter:
​http://www.pas.rochester.edu/~shuleli/collapse/light5.gif
normal ambient line cut:
​http://www.pas.rochester.edu/~shuleli/collapse/light10.gif

currently running the collapse setup with wind running through.

Adam,

Three people have expressed some interest in AstroBEAR: Riccardo, Radha and Suxing. The main issue is whether anybody would have time to work with a new code. That might be influenced by the level of support that you can supply and if we can find a student willing to work on these problems.

Riccardo would want to try a variation of the problem that you were originally looking at (presumably with a corrected treatment of the boundary conditions). Apparently no additional physics modules would be required. His comments were:

I am interested in this code if I can find at least one problem that is very well defined, doesn't require upgrades to the code and for which our codes cannot provide an answer. I think I may have found it, 3D ablative RT in a slab with gravity. That could tell us if there is a major differentce in 3D multimode ablative RT with respect to 2D. It uses the same input I already gave to Adam years ago.

Radha and Suxing might be interested to use AstroBEAR to look at other problems that would require developing some additional physics modules. A description of these is in the attachment.

---

Possible physics issues that can be addressed with AstroBEAR Nov 9, 2012

16. B. Radha and S. X. Hu

Problem classes fall under either short wavelength studies or 3D studies. Problems requiring some form of laser energy deposition, such as study of imprint or evolution of target defects are not listed below.

1. Short wavelength mixing at plastic (CH) gas (deuterium or deuterium/tritium etc.) interfaces for different Atwood numbers and materials with different atomic numbers, during the deceleration phase of implosion and hot-spot formation.

Physics issue: How significant is the classical interface for determining target performance (fusion yields)? Physics needed:

1. Implicit heat conduction – Can we get away with explicit if we limit time steps (probably not)
2. Multigroup diffusive radiation transport – Radiation plays an important role in setting up the density profile and the growth of Rayleigh –Taylor at the related unstable interface influences the growth at the material interface.

Scope:

1. This problem is not accessible by HYDRA.
2. While radiation is not critical to initial studies, they need to be added eventually to realistically model plastic shell implosions.
3. How do we know that numerical noise is not overwhelming the problem?

2. 3D assembly of hotspot

Physics issue: How does 3D physics influence the assembly of the hotspot in cryogenic DT-implosions? Physics needed:

1. Implicit heat conduction – Can we get away with explicit if we limit time steps (probably not)
2. Alpha transport

Scope:

1. This problem can be tackled by HYDRA though the effort at Livermore seems to be more toward modeling the integrated implosions as opposed to trying to understand the basis of the physics.
2. Where will the computational resources come from?

3. The effect of magnetic fields in implosions (3D problem)

Physics issue: Do magnetic fields influence heat conduction in the compressed core – particularly for Polar Drive implosions where we see significant variations in the symmetry of the hot spot in OMEGA implosions.

Physics needed:

1. Implicit heat conduction – Can we get away with explicit if we limit time steps (probably not)
2. Self-generated magnetic fields.
3. Effect of magnetic fields on heat conduction coefficients.

Scope:

1. This problem is likely unique to LLE and Polar Drive.
2. Where will the computational resources come from?

4. Studying the magnetic field reconnection in laboratory astrophysics NLUF experiments on OMEGA (this could be a possible application in not-so-near future)

Physics issue: How magnetic field reconnection starts and how magnetic energy is released as kinetic energy of particles?

Physics needed:

1. Self-generated and/or external B-field evolution
2. Multigroup diffusive radiation transport

Scope:

1. This is one of the grand challenges in astrophysics and is unrelated to ICF. It belongs to the basic science category. However, it may relate to LLE through reconnection experiments being conducted on EP.
2. The 3D nature might be crucial.

Last week was spent out of town and so not much to share. Today, focus has been on reading Toro. I would like to move through this task of reading and coding in the next month. Oh, and as far as holiday traveling, I will be out of town beginning tomorrow until next Tuesday, visiting relatives in St. Louis, Missouri.

Happy Holidays to everyone.

• Exchanged some emails with Francisco, and now I have some jet parameters to try out. It seemed a little odd to me that they used this aluminum cooling table (Post, et al 1977) to estimate their cooling parameter, but in simulations using Gorgon they used something else. I will give this Al cooling table a try, and see what we get.
• Read the mach stem thesis. Completed half of the initial mach stem runs, and created a new page for documenting everything on this project: MachStems. Hopefully, Kris is ready to start looking at these.
• Met with Jonathan about implementing cylindrical coordinates. I expect to make significant progress on this within the next month or so.
• All that token nonsense is done on my end, just waiting for activation now.
• Have not heard from Pat or Anna in a while about the radiative shocks.

• Golden Version & Blue Streak
  1. Found two modules failed on blue streak (ticket #265, #266)
  2. AstroBEAR running slow on Blue streak (see Jonathan's post on Blue streak performance)
  3. Queue of Blue streak doesn't work properly and the system seems unstable

• Tickets
  1. New: #265 (GravitationalCascade failed on blue streak), #266 (IsotropicTurbulence terminated on blue streak), #267 (Issues with BlueStreak and optimization above O3)
  2. Start working on #264 (Parallel IO)

• Outside Users
  1. Yat-Tien made some progress running the code. He will post the blog and call-in meeting next Monday (11/26)
  2. Shazrene asked to download the code. Account on local machines for her?
  3. Download page?

Runs:

• 1 Jupiter mass almost complete. ​movie here
  
• 10 Jupiter mass running.​movie here
  

• 20 and 5 Jupiter mass stopped because of a bug

​bug1
​bug2

what could be the cause ?

• affects multiple runs
• can happen before or after a restart
• golden version
• happens in different parts of the box
• doesn't seem to be a 'stepping on the boundary' kind of bug since it should be a little more symmetric.

Any ideas ?

• Looked at BlueStreak performance
• Made new figures for ​paper
• Working on Christina's run #267
• Preparing talk for Berkeley on 12/7…

We are running our production sims on Kraken and Ranger.
They will likely to be complete after the break. In the meantime, I worked on getting visit to run in parallel on Kraken and we now have some partial frames from the first run, that is, 1 Jupiter mass accretion disk.

Movie (0 → 1.5 orbits): ​here

The patterns we noticed in the previous sims are now more visible. Below are stills of frames 4 to 7, (for color legend look at the movie linked above).

Density profile (horizontal slice).

This run definitely features heavy mixing between the cold disk material and the hot ambient.

Temperature profile (vertical slice).

TOKES ?

CRL 618

Jet sim running.	New clump sim running

AGB r^-2 radiation pressure force

Running tests on the global binary sims, i.e. 2 stars in the grid, 40 AU.

We talked about including a A* r^-2 force term to the momentum equation to account for the wind's acceleration region. We need to decide on the strength of A.

the mom equation is:

v dv/dr = - 1/rho dp/dr - G m r^-2+A r^-2…(1)

it's useful to consider the relative forces: Gamma = A/Gm. So far we've used Gamma=1, but we want 0<Gamma<1.

Up: wind velocity / Mach. Bottom: wind density. VS. distance and Gamma values from Intro. to stellar winds, Lamers & Cassinelli.

Polodial magnetic collimation

This is essentially the same flow structure we find for strong fields in the lab, where the cavity is small and the plasma is immediately collimated into a jet, very close to the source (see panel 3 of Fig 4. in the article I sent you). Few things worth trying: (i) The interesting part of the flow in your case, seems to be under-resolved. Most of the computational box is not focusing on the interesting dynamics. (ii) Field lines not initially in the wind slide radially in the equatorial plane. However, if they are anchored in a disk the foot points should not move. This should give you field lines looking more like the ones on the inside of the cavity. If they are not anchored, then this is fine. (iii) You can lower the magnetic field or increase the wind mass loss rate, and should still get something like what you have now, but with a more elongated cavity.

Andrea: On the inside your flows look much more like the laboratory ones, here is an image from your simulations (colour map is density, vector is velocity, and lines is magnetic field lines).

AGN jet trucnation

Dark spot in the weak AGN radio jet survives for 30kyr when the red giant crosses its axis!

I ran a fixed grid colliding flow setup - with self gravity and cooling etc… for a few time steps to compare the performance of bluestreak when using 1, 8, or 16 cores on a single node as well as a single core on grass. I also compared different levels of optimization (-O3 and -O4). I tried -O5 but the code dies soon after outputting the first frame. I have tried compiling with traceback enabled to find out where the code is dying - but I don't seem able to get any useful information from the core dumps. I'm not sure what is going wrong there.

The figure below shows the cell updates/core/second over the first 10 timesteps for various numbers of cores on bluestreak - and for optimization levels 3 and 4 - as well as a single core run on grass with level 3 optimization (1_3g).

In any event the performance is about what you would expect. The code seems to slow down after a few time steps due to either CPU temperature - or perhaps cooling source terms… Optimization does not make much of a difference < 1% (at least going from O3 to O4). Also in going from 1 to 16 cores per node - the efficiency only drops around 10% which is surprisingly good. Overall the speed per core is about 25% that of grass. Given that grass has a 2.5 GHz processor and that each core on BlueStreak is supposedly running at 1.6 GHz - this is a little surprising. A single bluestreak core appears to be performing at 39% of what you would expect taking into account CPU speed.

Last week of work for me (Ivan).

A little summary of what we have done:

• We begun with a disk embedded in a uniform and hot ambient
• Material infall from the ambient would invalidate our initial conditions
• Developed a HSE solution to keep the ambient stable
• Put the disk back in the stable ambient
• Got acceptable results
• Now waiting for production runs to finish

Agenda for the week:

• Add modules used for runs to AstroBEAR repo (for future reference or work)
• Get production runs done (it is just a matter of waiting).
  

BE_HSE:
Should be picked up and developed by someone, it looks promising.



Parallel HDF5:
see blog post ​here

triggered star formation Finished some runs for pure BE sphere with Boss setup without shock to see how the BE sphere does. Surprisingly all of them expands. The original Boss 2010 setup has a cut-off on BE profile at 0.058 pc, the density at the outer edge of the cloud is 10 times denser than the ambient. Using the shock crossing time as the time scale, we see the following expansion:
​http://www.pas.rochester.edu/~shuleli/clump_paper/cut_small.gif

Without cut-off, we can zoom out and let the density of the outer edge relax to the ambient value, resulting in a much larger cloud, the cloud can then hold up better, although still expands:
​http://www.pas.rochester.edu/~shuleli/clump_paper/be_large.gif

Or we set the ambient density to be higher than what's suggested in Boss' paper:
​http://www.pas.rochester.edu/~shuleli/clump_paper/zoom_linear.gif

In their 2010 paper, they stated that for 2.5D isothermal, the BE sphere "does not collapse, but instead oscillates around its initial equilibrium structure, over a time period of at least 1e+6 yr". 1e+6 yr in our simulation is about 10 shock crossing time in our simulations. I'm running a simulation with a wind the same as in the 2010 paper and cloud the same as in the first movie.

resistive clumps Finished some of the unfinished strong magnetic field runs (beta = 1). Now we have weakly or strongly magnetized, parallel or perpendicular with Reynolds number 100, 1000 and infinity. It seems I also have data for weak parallel Reynolds number = 10 run done on kraken, but I haven't got time to look at data. I've done the visualizations for all the clump density plot. As an example, see beta = 4 density comparison at 2.5 crushing time:
​http://www.pas.rochester.edu/~shuleli/clump_paper/Resistive_MHDClump.png

papers I revised the previous paper and referee response according to Adam's comments. I'll give it a final read tonight and send it to Adam and Eric.

• Golden Version

1. checked in 1162:851b4b9a604e with Copyright — thank Ivan

!    Copyright (C) 2003-2012 Department of Physics and Astronomy,
!                            University of Rochester,
!                            Rochester, NY
!
!    global_declarations.f90 is part of AstroBEAR.
!
!    AstroBEAR is free software: you can redistribute it and/or modify    
!    it under the terms of the GNU General Public License as published by 
!    the Free Software Foundation, either version 3 of the License, or    
!    (at your option) any later version.
!
!    AstroBEAR is distributed in the hope that it will be useful, 
!    but WITHOUT ANY WARRANTY; without even the implied warranty of
!    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
!    GNU General Public License for more details.
!
!    You should have received a copy of the GNU General Public License
!    along with AstroBEAR.  If not, see <http://www.gnu.org/licenses/>.

2. Outside users: Created wiki and local machine account for Scott lucchini, created wiki account for Shazrene Mohamed. IO: waiting for the result so that we can start posting the beta version result in a blog. Attended LLE meeting.

• Blue streak

1. Currently has some problems with the Queue system. won't do a update until after Thanksgiving.
2. Post-processing with plots but one works: ​https://clover.pas.rochester.edu/trac/astrobear/wiki/u/MyCurentTests
3. Working on slow-cell-updating

• Tickets:

1. New #264 (Parallel hdf5 IO)

• got my NICS token in the mail, still need to activate it
• read up on cylindrical coordinates in Athena, ​http://adsabs.harvard.edu/abs/2010ApJS..188..290S. Now I need to find and alter the applicable routines in astrobear.
• I am running the mach stem simulations now, at a higher resolution. Images/movies will be posted soon. I did all the analytic calculations, so I have a set of runs planned.

I was thinking I'd do for each gamma a run at Rcrit, and a run ± 0.2 Rcrit…so with 4 different gammas, that's a total of 12 runs.

gamma	Rcrit/Rclump
1.6667	1.4679
1.4	1.3932
1.2	1.3229
1.1	1.2766

Relating this to the real situation of two identical clumps side by side is simple. The distance to the reflecting wall is half of the separation distance if we take the separation distance to be the distance between the clump centers. So the critical separation distance is twice Rcrit.

• I will also start setting up my module to do the problem with interacting bow shocks where one is moving.

Why ?
Serial HDF5 can be slow:

• when many processes try to write/read to same file
• when handling large files
• when a large amount of IO is done (postprocessing…)

How AstroBEAR handles IO:
Master queries all processes asking for data, data is then written to chombo file.

Original image from ​http://www.astro.sunysb.edu/mzingale/io_tutorial/

Simplified version of astrobear IO routine:

offset = 0
for every active process
   fectch data
   write data to chombo beginning at offset
   update offset

How this would be done in parallel:
Every process would write at a different offset.

Original image from ​http://www.astro.sunysb.edu/mzingale/io_tutorial/

Requirements of parallel hdf5 usage:

• MPI-IO (should be on all machines)
• Parallel filesystem (not on all machines)

PHDF5 should be considered as an addition, not a replacement for serial HDF5.

Parallel file accessing aspects:
All processes that have access to a file must do basic operations collectively:

• Open/close/create file
• Open/close/extend datasets

Other operations can be done either collectively or independently:

• File read
• File writes

The easy part:

• The two libraries are incredibly similar.
• We can adapt our serial routines to use parallel, not implement parallel IO from scratch.

The challenging part:

• Every process must know where to write.
  
• An algorithm is needed.
  
• Concept explained in image below

Poloidal collimation New run finished. Andrea's looking at the data. New runs to be ran soon.

AGN jet truncation Run with the star at the jet's axis is running:

CRL 618, new run (v_clump=250km/s, higher densities [stringer cooling]) running in Bgene:

Binary. Debugging old module to use the golden version.

hedal procs writing answers.

Random thought of the day:

Maybe for new releases of the code, we can use different animal names, i.e.:

AstroLION, AstroCRANE, AstroSHARK,….

• Worked on Spectral Prolongation
• Using ​BlueStreak for post-processing of colliding flow runs
• Created ticket #264 for parallel IO
• Setting Up Christina's MHD Colliding FLow run on ​BlueStreak

Production runs:

• Currently running on BlueStreak
• Had 4 models initially (20 M_J,10 M_J,5 M_J,1 M_J,)
• 1 Jupiter Mass Disk sim was stopped at ~2 orbits (it was almost completely disrupted).
  
• We now have a nice intermediate case (5 M_j)
  

Initial conditions:

~ 1 Orbit

~ 2 Orbits

… While we wait for the production runs to advance.

BE sphere:
Mini movie of a .25 M_J BE sphere orbiting around hydrostatic star core:
The sphere is within the shredding radius.

Golden version:
Tagging question. Copyright year ?
X-2012

AGN jet truncation. Data with the desired resolution:

Now running the long term evolution of this run and the on-axis version.

Kraken. Waiting for reply from teragrid on the tokens.

CRL 618. Phone call with Bruce. The numerical section of the paper is in our hands. He thinks that the other sections of the paper are in very good shape. New runs:

• v_j=200km/s with higher dens_j and dens_amb to produce thinner lobes. I addition, we will add a tracer field that follows the onion-like density distribution of the ambient, to see if molecular H in the ambient can get "glued" to the lobes. This is motivated by the effect that the ambient rings have on the structure of the lobes:

.

Global binary sims (the 2 stars in the grid). Test runs have begun.

Poloidal magnetic collimation. Still working on getting the parameters that we want.

PN luminosities vs. Mdot. Met with Eric B. and chatted about his student, Scott, and the project about PN jets with different Mdots. Will meet with Scott this week.

• still waiting to hear from Anna (rad shock emissions), Francisco (cooling jets), and now Kris (mach stems)
• did some analytic work on the mach stem project following Kris' page: ClumpClump/MachStems

My problem set up will now allow me to test different gammas and different separation distances. ehansen10312012

• going to start working on implementing cylindrical reconstruction for 2.5D ehansen10202012

• Golden Version & Blue Streak
  1. tested and checked in 1156:86ff14bc8662 and 1161:f528b4a4b297 .
  2. Ivan found a easier way to handle the preprocessor issue on blue streak, Now the makefile.bluestreak and *.F files are deprecated in rev 1161:f528b4a4b297. global_declaration.f90 with Absolute path is deprecated also. So compiling and running AstroBEAR on blue streak is just as blue gene/P now. (Ticket #245)
  3. Tried installing mercurial on blue streak
  4. Copyright year for Golden Version AstroBEAR
  5. Ivan and Jonathan are running jobs on blue streak.

• Tickets
  1. New: none
  2. Closed: #251(wiki error), #262(Big/Little Endian on blue streak)