DSL Designing And Evaluating For Ocean Models

The development of ocean models requires knowledge from different domains. One aspect of the modeling is the model configuration that takes place in code files or parameter lists. The process of configuration of each ocean model is different and their users must know the differences. To make a configuration of the ocean models is easy we can implement a DSL that generates valid configuration files for each model. In this thesis we design and implement a such configuration DSL. Hereby we study the use cases scenarios involving model parameterization and one ocean model. Based on the findings we designed and implemented the DSL. Although the DSL does not generate all configuration files, the evaluation shows that the concept works.

  1. Serafim Simonov. 2020. DSL Designing And Evaluating For Ocean Models. Kiel University. Retrieved from http://eprints.uni-kiel.de/51160/

First Visualization of the UVic Architecture

Our goal is to understand the composition of climate and ocean models to support their modularization and future development. Recently, we applied runtime monitoring on the MITgcm model. Based on our experience there, we applied the technique to the Earth System Climate Model (ESCM) of University of Victoria, Canada. Please be aware that these are very early results and may be erroneous.

The UVic model can be compiled with GNU Fortran (gfortran), but the current setup, we used, only produces a running executable with the Intel Fortran compiler (ifort). Fortunately, ifort support the same interface for runtime instrumentation as gfortran. Thus, we could apply the same probes in this context.

Based on this setup, we recorded 79 GB of binary monitoring data from a partial model run. We aim to have a complete run in future, but for the proof of concept, a partial run is sufficient. For our analysis we aimed to use the standard Kieker trace-analysis tool.

However, the Kieker trace-analysis tool uses call traces to reconstruct the deployed architecture. It is designed that way based on knowledge from web-based and service-oriented services. They have usually a small set of calls in a trace, triggered by an incoming event, message or request. In models, this is quite different. They are called once and run for a long time. Essentially, this results in one big trace. In our case 79 GB trace. This would not fit into memory, and even if, it would be very slow to process. Thus, we created a new architecture reconstruction tool based on another set of Kieker analysis stages. Utilizing this tool, we could generate our first component and operations graphs. The first component graph can be seen below.

UVic architecture based on Kieker monitoring data. Files are considered to be components.

We will continue our analysis to provide better readable graphs.

Thematic Analysis Tool

Thematic Analysis is a method to analyze text, audio, video and other material qualitatively. In OceanDSL, we use this method to analyze interview transcripts. Goal of the thesis is to develop suitable tooling or parts of it and evaluate the tool based on existing transcripts.

Type Bachelor / Master

Task Create a coding editor in Eclipse supporting coding of text, tagging/categorizing, code and category editing.

Task Provide an interactive visualization based on ELK/Kieler.

Task Develop a web-based visualization tool

Features

  • coding, categorizing/themes
  • regrouping of codes
  • re-coding
  • splitting codes
  • merging codes

Extend the Design and Evaluate a Configuration DSL for Ocean Models

Type Bachelor (evaluation with MITgcm or UVic scenarios)

Type Master (evaluation with MITgcm and UVic scenarios and in conjunction with domain experts)

Task Based on a given set of domain knowledge, concepts and sample configuration scripts, design a textual, external DSL to control configuration of models based on the two case studies MITgcm and UVic addressing all three deployment scenarios (local, dedicated host/node, and kubernetes). The configuration language will only address static configuration pattern comparable to the Sprat setup language. It might be possible to derive the OceanDSL configuration language from the Sprat setup language. The DSL includes a code generator usable standalone and within Juypter for the DSL.

Note This work will be based on a already existing DSL which will be extended.

Resources

  • The Sprat Approach (see also below), specifically the simulation configuration DSL
    • Sprat http://eprints.uni-kiel.de/32070/
      • Chapter 7 Especially 7-7.2.2
      • Metamodel in 7.2.3 to understand the general relationship of the terminology, the property definition in 7.2.3 is done in Object-Z, but essentially the name on the top refers to the class from the metamodel and the pairs of names in the inner frame are the properties for the class and additional constraints. Inheritance in Object-Z is shown in the defition of Internal_DSL and External_DSL (the internal DSL has also a property hostLanguage).
    • 7.3 Applying the Sprat ApproachTo understand how and where the separation between domains happen, you need to define roles. This is described in 7.3.1
    • Chapter 8
      • 8.3 The Sprat Ecosystems DSL
    • Source Code
    • https://github.com/cau-se/sprat-ecosystem-dsl-xtext
  • Introductions to how to write a thesis https://www.se.informatik.uni-kiel.de/en/student-theses/useful-hints
  • Case Studies
  • External DSL notation for syntax and semantics

Design and Evaluate a Deployment DSL for Ocean Models

Type Bachelor (evaluation with MITgcm or UVic scenarios)

Type Master (evaluation with MITgcm and UVic scenarios and in conjunction with domain experts)

Tasks Based on a given set of domain knowledge, concepts and sample deployment scripts, design a textual, external DSL to control deployment of models based on the two case studies UVic and MITgcm addressing all three deployment scenarios (local, dedicated host/node, and kubernetes). The DSL includes a code generator, interpreter or Jupyter kernel which performs the deployment.

Resources and Notes

  • Ansible is a deployment and configuration language https://www.ansible.com/overview/how-ansible-works
  • The deployment process for ocean models can be quite different from those established in enterprise software where Ansible is designed for. Typical processes is (so far) configure code, compile, configure program, setup, run. All these tests can be done locally or remote. However, at the point when it goes remote it stays remote from that point on. That means when code configuration happens locally, but compiling is done remote, the rest is also done remote.
  • Access to remote machines happen via ssh (legacy) or via Jupyter (latest).
  • Version management and file synchronization is done via git.
  • The DSL must be able to describe all phases of the deployment. This could be done similarly to a build pipeline from
  • Sprat Deployment based on Ansible
  • The Sprat Approach (see also below), specifically the simulation configuration DSL
    • Sprat http://eprints.uni-kiel.de/32070/
      • Chapter 7 Especially 7-7.2.2
      • Metamodel in 7.2.3 to understand the general relationship of the terminology, the property definition in 7.2.3 is done in Object-Z, but essentially the name on the top refers to the class from the metamodel and the pairs of names in the inner frame are the properties for the class and additional constraints. Inheritance in Object-Z is shown in the defition of Internal_DSL and External_DSL (the internal DSL has also a property host language).
      • 7.3 Applying the Sprat Approach To understand how and where the separation between domains happen, you need to define roles. This is described in 7.3.1
    • Source Code
    • https://github.com/cau-se/sprat-ecosystem-dsl-xtext
  • Introductions to how to write a thesis https://www.se.informatik.uni-kiel.de/en/student-theses/useful-hints
  • Case Studies
  • External DSL notation for syntax and semantics Syntax EBNF

Identify, adapt and develop Code Analysis Tooling for Fortran, C and Python

Type Bachelor (one language, one analysis, limited survey)

Type Master (full survey including preprocessing, one analysis)

Task Create a survey for code analysis tools including style checkers and components for Fortran, C and typical preprocessors. Components for such technology are parsers, lexers, ASTs.

Sources & Notes

  • Starting point: Existing Fortran grammars and tooling
    • ROSE project
    • Open Fortran Grammar

Complexity Analysis of Ocean Models

Type Master

Task Analyze existing climate and ocean models to identify internal dependencies and the architecture. Subsequently, identify code which is not used (dead code).

Potential models to analyze:

  • MITgcm, UVic (available)
  • NEMO, ICON, ECHAM5/6 (future candidates)

Key questions

  • How to identify internal dependencies/discover the architecture?
  • How to identify dead code in a model?
  • Determine complexity

Tools

  • Static code analysis utilizing grammar and preprocessor stuff (maybe Sergej and Ralf)
  • Kieker 4 C runtime architecture
  • Joint visualization with new Kieker filters via dot and graphml

Unified access to scheduling systems

Type Bachelor

Task Analyze different schedulers, like Slurm, NQSV regarding their features.

  • Research which batch systems / schedulers are in use in specific HPC installations
  • Identify their features in a feature matrix
  • Identify common concepts and/or an abstraction the functionality
  • Ask modellers regarding the use of the features (questionnaire)
  • Sketch a basic API, internal (python)/external DSL usable within Jupyter

Sources & Notes

Architecture Analysis of Climate Models based on Kieker Runtime Data

We are analyzing the architecture of climate and ocean models utilizing runtime monitoring data reflecting call traces. Currently, we are testing our technical approach based on the MITgcm and its set of prepared verification experiments including all tutorial setup.

Today we analyzed two experiments, namely tutorial_barotropic_gyre and tutorial_global_oce_biogeo. For the former, we derived components based on files. For the latter, we derived components based on the package and source code directory structure, namely, each MITgcm package represents one component and the main code block is represented by the BASE component.

Preliminary results can be seen in the following two figures:

Components based on files

As you can see, the file based components are quite numerous and result in a fast graph. In contrast the, package based component graph looks as follows.

Components based on packages and main source tree

In the latter graphic you can see cyclic dependencies between BASE and most components. This might be the result that BASE comprises of all non-package code files. Thus, a better separation in this area might be helpful. Still, the figure is much concise and understandable.

Due to technical issues, we cannot provide operation based images for both cases. There might be an error in the dot renderer.

Configuration and Parameterization DSL

Over the summer, we developed our first language prototype of a Configuration and Parameterization DSL. The DSL allows to specify parameters for the MIT General Circulation Model (MITgcm). The current prototype is limited to support configuration for the following tutorial examples:

Currently, we extend the DSL to be able to support all tutorial examples and features of MITgcm and prepare the support of the UVic model.

Project Resources