Improving performance and debugging of transactional applications through programmer annotations

• Grant number: 16/12103-7
• Support Opportunities: Scholarships in Brazil - Master
• Start date: November 01, 2016
• End date: July 31, 2018
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Alexandro José Baldassin
• Grantee: Bruno Chinelato Honorio
• Host Institution: Instituto de Geociências e Ciências Exatas (IGCE). Universidade Estadual Paulista (UNESP). Campus de Rio Claro. Rio Claro , SP, Brazil

Abstract

Transactional Memory (TM) allows programmers to use the concept of transaction while writing concurrent code. In this context, a transaction may be understood as a block of instructions that is executed atomically and in isolation, that is, each transaction intermediate processing states are not visible to the others. Although initially confined to the academic environment, TM is quickly becoming popular. One reason is the addition of transactional hardware in the new processors released by Intel and IBM, as well as the compiler support provided by the GNU C Compiler (GCC). The big advantage of the transactional model lies in its high abstraction level, making it easier for programmers to write parallel software while avoiding classical synchronization errors caused by locks, such as deadlock. Unfortunately, the software support for transactional execution still does not provide high performance. In particular, the transactional code, generated by compilers and the runtime system, is still considered somewhat inefficient. In order to tackle this problem, this Master's project aims at creating methods based on programmer annotations in order to aid the compiler in the generation of more efficient code. This additional knowledge provided by programmers seeks to avoid over instrumentation usually generated by compilers in order to guarantee correctness. At the same time, these annotations can help the compiler and the runtime system to provide feedback so as to help programmers to identify errors while writing transactional code. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

HONORIO, BRUNO CHINELATO; DE CARVALHO, JOAO P. L.; BALDASSIN, ALEXANDRO; BIANCHINI, CD; DESOUZA, PSL; DEBARROS, COF; FERREIRA, RAC. On the Efficiency of Transactional Code Generation: A GCC Case Study. 2018 SYMPOSIUM ON HIGH PERFORMANCE COMPUTING SYSTEMS (WSCAD 2018), v. N/A, p. 7-pg., 2018-01-01. (16/12103-7, 16/15337-9)

Academic Publications

(References retrieved automatically from State of São Paulo Research Institutions)

HONORIO, Bruno Chinelato. Improving the performance of transactional applications through annotations made by the programmer. 2018. Master's Dissertation - Universidade Estadual Paulista (Unesp). Instituto de Biociências Letras e Ciências Exatas. São José do Rio Preto São José do Rio Preto.