THAP015

Title: THAP015

Research Question: How can Data Access, an object-oriented library, be implemented effectively on various platforms to improve EPICS performance and memory usage?

Methodology: The research team tested Data Access on several platforms using different compilers, including Pentium PCs running Linux, Sun Ultra-30s running Solaris, and Motorola 68k running vxWorks. They used a small test application to measure the performance of assigning between containers of different data types.

Results: The tests showed that Data Access can be implemented on various platforms using different compilers. The library offers several advantages over the existing data conversion library, such as extensibility, range checks, type safety, and support for multi-dimensional arrays. However, the team found that certain advanced features of C++ had to be avoided or replaced by workarounds due to differences in compiler implementations.

Implications: The implementation of Data Access on various platforms using different compilers demonstrates its potential to improve EPICS performance and memory usage. The test results provide a first impression of the library's behavior, but further tests on a wider set of platforms with mature compilers are necessary for a complete understanding of its performance and efficiency.

Link to Article: https://arxiv.org/abs/0111036v1 Authors: arXiv ID: 0111036v1