An instruction set extension for java bytecodes translation acceleration

Java has become popular in a wide range of applications. Just-in-time translation is crucial for providing efficient execution of Java programs. This paper presents an architecture extension to RISC processors that accelerates Java bytecodes translation. Our results show that the incorporation of this technique in a 4-way superscalar RISC and in one high performance embedded processor gives an average speedup of 2.95x and 2.29x respectively. A first order approximation using ASIC synthesis tools shows that this acceleration is performed with only a small increase in hardware.

[1]  J. Michael O'Connor,et al.  picoJava-I: the Java virtual machine in hardware , 1997, IEEE Micro.

[2]  George Economakos,et al.  A cache based stack folding technique for high performance Java processors , 2006, JTRES '06.

[3]  Ramesh Radhakrishnan,et al.  Allowing for ILP in an embedded Java processor , 2000, Proceedings of 27th International Symposium on Computer Architecture (IEEE Cat. No.RS00201).

[4]  Kin F. Li,et al.  A robust stack folding approach for Java processors: an operand extraction-based algorithm , 2001, J. Syst. Archit..

[5]  Robert Wilson,et al.  Compiling Java just in time , 1997, IEEE Micro.

[6]  Han-Jong Kim,et al.  Sim-ARM1136: A case study on the accuracy of the cycle-accurate simulator , 2006, Microprocess. Microsystems.