SPEC2000 Daily Testing

• These results are not reportable according to SPEC rules.
• Less than the minimum number of runs have been used.
• Some tests may not have compiled or executed properly.
• The goal of these daily runs is to monitor the performance of generated code using standard optimization flags. Therefore, these scores are probably lower than what could be obtained using the most aggressive optimization options.
• From time to time, I may change the compiler flags used to compile the benchmarks. If you notice sudden peaks or depressions in the graphs, check the log files for those dates to make sure that they are not caused by the new flags. Ideally, this should be automatically marked by the scripts, but I haven't gotten around to implementing this feature yet.

1. Baseline compilers.
• gcc version 3.1.
• gcc version 3.2 20020903 (Red Hat Linux 8.0 3.2-7).
• gcc version 3.3.2.
• gcc version 3.4.0.
• Intel(R) C++ Compiler for 32-bit applications, Version 7.0 Pre-Beta Build 020415Z.
• gcc version 3.4-llvm 20030924 (experimental).
2. Global SPEC results.
• Scores.
• Build times.
• Run times.
3. Individual SPEC results.
• Scores.
• Build times.
• Run times.
• Executable sizes.
4. Bootstrap and size stats for GCC.
5. Daily log files, diffs and SPEC reports.

Return to index page.