Titanium Language Reference Manual (Version 2.20)

Abstract

The Titanium language is a Java dialect for high-performance parallel scientific computing. Titanium's differences from Java include multi-dimensional arrays, an explicitly parallel SPMD model of computation with a global address space, a form of value class, and zone-based memory management. This reference manual describes the differences between Titanium and Java.

Macro modeling of liquid crystal cell using machine learning method: reservoir computing approach

“…We implemented our machine learning model into a commercial simulator SmartSpice 29) with Verilog-A language (Version 2.3.1). 24) 3. Results To run the macro model at high speed in the circuit simulator, it is desirable to keep the dimension of the reservoir vector (Dr.), which affects the computational cost, as small as possible.…”

“…We adopted a nonlinear dynamical equation in continuous time 22) for updating reservoir states and discretized it using the Runge-Kutta method, 23) allowing the prediction at arbitrary time steps as determined by the circuit simulator. We implemented the model into a circuit simulator by using Verilog-A language 24) and verified that it worked well in the pixel circuit design. Furthermore, we discussed how parameters that control the length and the update frequency of memory in the RC model, which are important issues when dealing with macroscopic characteristics, affect prediction accuracy.…”

A Higher-Level and Portable Approach for GPGPU Programming

“…In the language, we attempted to combine features of C# language together with array programming primitives which would allow higher-level GPGPU programming. Such an approach is quite common for modern parallel programming languages, such as Titanium [20] or Chapel [21], so we follow it. Technically, our language is based on .NET, and array and functional types are translated into .NET interfaces.…”

Textual alignment in SPMD programs

“…A type system is designed to determinate which branches can produce distinct behaviors and, in this case, compare the numbers of operations [1,11]. This work has been used for the design of the Titanium language [23,6]. A later proposal introduced textually aligned barriers in Titanium by revisiting structural correctness.…”