Update 1/27

Radiation pressure paper

Revision submitted. Hopefully it'll be accepted soon and I can move the data off BlueHive.

MHD

Ramp vs No-Ramp

http://www.pas.rochester.edu/~adebrech/MHD/Run2_ramp_comp.png

Slight differences. But the non-ramping version is working, so that's probably the better way to go. Working on some resolution tests now.

Charge Exchange

http://www.pas.rochester.edu/~adebrech/ChargeExchange/HD209458b_charge_exchange0000.png

I have 10 more frames of this, but no room on BlueHive currently. The low-pressure wind is working (have 5 frames or so), but I believe I've pretty much used up my allotted portion of the allocation on Stampede2.

I will start the postprocessing on BlueHive, so that we can start looking at synthetic observations.

AMR line transfer

This may become useful for the higher field cases. In order to decrease the discontinuity between the planet and the ambient, we need higher resolution around the planet - but we can force the rest of the simulation to stay at e.g. level 4 maximum while refining the planet to a much higher level. However, this would be practically impossible to run in the "projected to max level" version of the line transfer routine. I'm working on updating and merging this into all of my branches.

Currently waiting on a timing test for the buffered nonblocking send with ray stack version, to see if that's an improvement over the last looping version (which may or may not have used IBsend during the test - if this isn't an improvement, probably have to check again).

At some point I would like to make a graph of the dependencies and see if there's a best way to prioritize the rays.

HPC summer school

Statements are due midnight tonight, AoE.

Exoplanets III

Talk abstract for MHD submitted. Deadline is this Friday, I assume we'll hear back Feb-Mar. Working on flight and hotel.

Misc

I've updated the development branch to use the MPI_F08 module, which allows type and bounds checking, makes iErr optional, and in general is both better and the only way to use MPI that's completely compatible with the Fortran standard. Sent a merge request on Gitlab.

I now have my own professional domain: ​alexcomputes.com (was tired of using my much less professional free domain on my CV).

Comments

No comments.