Clang

Clang

LLVM-based stochastic error propagation analysis of manually developed software components

Published in Stein Haugen, Anne Barros, Coen van Gulijk, Trond Kongsvik, Jan Erik Vinnem, Safety and Reliability – Safe Societies in a Changing World, 2018

A. Morozov, K. Janschek, Y. Zhou

Figure 5 shows a typical LLVM-based workflow. An input source code, e.g. a C-code or any other of more than 20 LLVM supported languages, can be compiled with clang into the LLVM IR. Using the LLVM API, an optimizer can be built that produces an optimized version of the compiled LLVM IR, which, in turn, can be further compiled with llc for x86, ARM, PowerPC or other platforms. The optimizer can be built with so-called analysis or transformation passes. An analysis pass does not change the input IR, but provides analytical information, for instance, counts the number of instructions or builds a call graph. In contrast, a transformation pass changes and optimizes the original LLVM IR.

How AD can help solve differential-algebraic equations

View Article

Journal Information

Published in Optimization Methods and Software, 2018

John D. Pryce, Nedialko S. Nedialkov, Guangning Tan, Xiao Li

We have applied the Daets Lagrangian facility to various systems, including the following examples. Performance tests are on a 2017 MacBook Pro laptop with a 4-core 2.2 GHz Intel processor running Mac OS X 10.11.6. The C++ compiler is clang++ version 8.0.0. All numerics are in C++double.

Explore chapters and articles related to this topic

LLVM-based stochastic error propagation analysis of manually developed software components

How AD can help solve differential-algebraic equations