Simulations and distributed computing working group (SIMS-WG)

Role

SIMS-WG assists in optimizing the JETSCAPE workflow such that it scales well inside High Performance Computing (HPC) environments.

Goal

The most computationally demading modules in the JETSCAPE framework are those associated with the dynamical evolution of the soft medium with which jets interact. The present goal of the SIMS-WG is to perform a statistical Bayesian analysis of the soft medium, constraining the various parameters present therein.

Modules describing the medium created in heavy-ion collisions

There are four modules present in the JETSCAPE framework (v.1.4 and greater) that are used to simulate the evolution of the medium:

  • TRENTO: provides the initial energy density profile to be evolved.
  • Free-Streaming: evolves the initial energy density profile as well as other components present in a fluid, such as flow, for a short amount of time during which it is valid. Thus it provides a non-zero initial flow and a non-trivial initial energy density to be evolved by a hydrodynamical simulation.
  • MUSIC: is the hydrodynamical simulations that the JETSCAPE SIMS-WG will be using for the majority of the evolution of the plasma created in heavy-ion collisions.
  • SMASH: evolves the medium after the hydrodynamical simulation stops being valid. Indeed, the degrees of freedom in SMASH are no longer those of hydrodynamics (e.g. flow) but rather individual hadrons, that can be detected experimentally. For more info about SMASH, please click the link to visit its webpage.

Trento

An example of the initial energy density profile for a simulating a single high-energy heavy-ion collision generated by TRENTO. There are millions of such collisions happening in actual experiment, and to simulate them on a collision per collision basis,  each energy density profile has a different distribution of the peaks and valleys. Also, the height of the peaks and the depth of the valleys varies from one collision to the next. What is shown on the left is just an example of a particular simulation of such a collision. 

Free-Streaming

The above TRENTO initial condition is evolved for some amount of time (here 1.16 fm/c, but that will be determined much better following our statistical Bayesian analysis) giving the energy density profile seen on the left. Notice how smooth the energy density distribution is after free-streaming, and also note that the maximum size of the highest peak has been reduced (by a factor of about 2 in this case) while the overall volume occupied has increased.

MUSIC

After this evolution of 1.16 fm/c, the hydrodynamical simulation MUSIC evolves this further to obtain the final temperature and flow profile. Unlike free-streaming, this evolution is not unique and depends on transport coefficients (e.g. shear and bulk viscosity) present in the evolution.  For example, different values of parameters for the temperature dependent shear viscosity can lead to various shapes of the temperature (or energy density) profiles. The figure on the left shows the temperature profile of the medium with a constant shear viscosity over entropy density ratio (also known as specific shear viscosity). That figure was obtained after 5.25 fm/s of hydrodynamical simulation using MUSIC, and zooms in on the temperature profile ranging from 160 to 166 MeV and assumes a constant specific shear viscosity (see Phys. Rev. C 014902 (2018) for details of this calculation).

If instead one uses a temperature dependent specific shear viscosity (see Phys. Rev. C 014902 (2018)  for details), then a different temperature profile can be obtained as illustrated in the adjacent figure, also obtained after 5.25 fm/c of evolution.

Members

Conveners

PhD Students

Faculty

External Members

Jean-François Paquet

(Duke University)

Gojko Vujanovic

(Wayne State University)

 

Lipei Du

(Ohio State University)

Derek Everett

(Ohio State University)

Matthew Heffernan

(McGill University)

Weiyao Ke

(Duke Univesity)

Steffen Bass

(Duke University)

Charles Gale

(McGill Unversity)

Abhijit Majumder

(Wayne State University)

Chun Shen

(Wayne State University)

Matthew Luzum

(University of São Paulo)