Review form for CSM – reviews 2003-09-22 -------------------------------------------------- Paper: Lal Geroge, Matthias Blume. Taming the IXP Network Processor -------------------------------------------------- Reviewer name: Peter Bunus -------------------------------------------------- Short summary: -------------------------------------------------- The paper present a novel programming language called Nova for writing network processing application for the Intel IXP200 family network processors.The language provides support for efficient optimal bank assignments including register spilling, and allocation of aggregates. The authors propose back-end based on integer-linear programming (ILP).Within a framework based on ILP formulation the bank assignments problem, coloring of aggregates in the bank assignment context and the management of variables in multiple locations are also addressed. The main contributions: -------------------------------------------------- - Development of the Nova language for programming IXP network processors - The formulation of the bank assignment, coloring of aggregates in conjunction with bank assignments and the management of variables in multiple locations as an integer-linear programming (ILP) problem. Merits and weaknesses: -------------------------------------------------- The problem seems good (although, I admit, I am not an expert at all) and the solution - promising. The paper also uses good running examples to explain the concepts in the paper. However, the comprehension of the paper is weakened by the fact that the authors refers several times to Appel and George work where maybe a short explanation might be more appropriate. Significance: 8 Originality: 8 Interest to a journal on programming languages and compiler technology: 5 Quality of experimental evaluation: 7 Overall organization: 7 Presentation (language and style): 9 Length appropriate: 10 References appropriate: 10 Overall evaluation: 8 Recommendation Accept/ weak accept / weak reject /reject: Accept Your confidence in you review (1=novice, 10=leading expert): 1 Comments to the authors: -------------------------------------------------- minor issues: - An acronym is used in the abstract (CPS) without being previously introduced. For an unaware reader, it might be not so obvious that the CPS stands for continuation passing style. - A “dot” is missing at the end of the description text corresponding to Figure 1. Suggestions for improvement (for the authors) -------------------------------------------------- An evaluation between the code generated by the compiler and some hand written assembly code.