COMPANY Home Contacts Customers Testimonials PRODUCTS Overview NULLSTONE for C NULLSTONE for CSharp NULLSTONE for Java Request Information SUPPORT Release Notes Download PGP Information Service Report Write Us INFORMATION Performance Results Glossary of Terms RELATED LINKS Compiler Connection Compiler Jobs

Glossary of Terms Bimodal Performance Behavior Each NULLSTONE performance test has been designed to exhibit Bimodal Performance Behavior. That is, the optimization being tested dominates the code in the test, and failure to perform the optimization being tested will significantly affect the run-time performance of the test. Compile Time Measured in seconds, the time to compile a test. Lower is better. Ideal Rate Measured in iterations per second, the rate at which the test can be executed after the optimization under test has been successfully optimized. The Ideal Rate for a test is dependent upon system features such as architecture, micro-architecture, cache, and memory size. NULLSTONE Rate Measured in iterations per second, the rate at which the test code fragment can be executed. Higher is better. NULLSTONE Ratio A percentage between 0% and 100% inclusive, the ratio of the NULLSTONE Rate and the Ideal Rate. A NULLSTONE Ratio of 100% is a perfect score, and indicates that the compiler performed the optimization being tested. Compilation failures and execution failures are reported as 0%. Higher is better. Performance Regression A decrease in the NULLSTONE Rate for a performance test when compared to a previous release of the same compiler. Sample Size The number of tests in a particular category. © 1990-2012 Nullstone Corporation. All Rights Reserved.