berczik / phi-GPU-mole

Parallel Hermite Integration on GPU (Direct N-body on MOLE-8.5 cluster)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

phi-GPU-mole

Parallel Hermite Integration on GPU (Direct N-body on MOLE-8.5 cluster)


This is a short readme file for the first public version of "phi-GPU" N-body program.

The phi-GPU-mole.tar.gz file contain the exact version of code which was used for the large set of test runs on the Mole-8.5 GPU cluster of the IPE CAS, Beijing, China around April 2011. The results are in the code paper:

https://ui.adsabs.harvard.edu/abs/2011hpc..conf....8B/abstract


This is a fully GPU (and only NVIDIA GPU + CUDA code). We start to develop this code together with Keigo Nitadori around the end of 2008.

Basically we fully rewrite our earlier MPI parallel phi-GRAPE ("C" code) code using the new "C++" standards and CUDA calls. The code was written "around" the Keigo Nitadori & Junichiro Makino earlier high order (Hermite 4th, 6th & 8th) CPU integrator:

  • K. Nitadori and J. Makino. Sixth- and eighth-order Hermite integrator for N-body simulations. New Astronomy, 13:498-507, Oct. 2008.

The code directly work with the GPU's. No need in any kind of external (like SAPPORO) libraries.

The code is distributed "as is". Please, if you use the code for any "productive" or even "test" runs cite our papers:


Code details in HPC conference:

https://ui.adsabs.harvard.edu/abs/2011hpc..conf....8B/abstract

High performance massively parallel direct N-body simulations on large GPU clusters.

Peter Berczik, Keigo Nitadori, Shiyan Zhong, Rainer Spurzem, Tsuyoshi Hamada, Xiaowei Wang Ingo Berentzen, Alexander Veles, Wei Ge

International conference on High Performance Computing, Kyiv, Ukraine, October 8-10, 2011.

First large N astrophysical results:

http://adsabs.harvard.edu/abs/2012ApJ...756...30K

Formation and Hardening of Supermassive Black Hole Binaries in Minor Mergers of Disk Galaxies.

Khan Fazeel Mahmood, Berentzen Ingo, Berczik Peter, Just Andreas, Mayer Lucio, Nitadori Keigo, Callegari Simone

The Astrophysical Journal, Volume 756, Issue 1, article id. 30, 10 pp. (2012).

In this papers (especially in the first one) we are in detail analize the performance of the different parts of the code.

In any other questions or want to report some problems with the code please send me the e-mail or contact me in any other way.

All the best, Peter Berczik.


Dr.Sci. Peter Berczik

Main Astronomical Observatory, National Academy of Sciences of Ukraine, MAO/NASU, 27 Akademika Zabolotnoho St., 03143 Kyiv, Ukraine

E-mail: berczik@mao.kiev.ua Office: +38 (044) 526-47-71 Mobil: +38 (099) 503-72-62 Skype: berczik-office


About

Parallel Hermite Integration on GPU (Direct N-body on MOLE-8.5 cluster)