Benchmarks ========== Using http://hg.koziarski.com/performance_tracker/. These were run on my Powerbook 1.67GHz... it was otherwise idle except for me chatting on IRC. Tests were run using the benchmarking.rb environment provided with the following changes: config.cache_classes = false config.action_controller.perform_caching = false Makes sense to only change these two things as caching is the only aspect of dev mode these patches take a swing at. BEFORE (dev mode, no patches) ----------------------------------------------------------------------- benchmarking 3 runs with options 100 -bm=default page request total stddev% r/s ms/r html 46.08638 4.2755 2.17 460.86 xml 48.53045 5.6099 2.06 485.30 all requests 94.61683 4.6590 2.11 473.08 AFTER (dev_mode_performace_fixes v0.3 - 17.5x faster) ----------------------------------------------------------------------- benchmarking 3 runs with options 100 -bm=default page request total stddev% r/s ms/r html 2.88391 2.9670 34.68 28.84 xml 2.78086 8.4500 35.96 27.81 all requests 5.66476 3.6506 35.31 28.32 AFTER (dev_mode_performace_fixes v0.2) ----------------------------------------------------------------------- benchmarking 3 runs with options 100 -bm=default page request total stddev% r/s ms/r html 11.41378 8.2382 8.76 114.14 xml 11.08206 2.6687 9.02 110.82 all requests 22.49584 5.4876 8.89 112.48 AFTER (with dev_mode_performance_fixes plugin v0.1) ----------------------------------------------------------------------- benchmarking 3 runs with options 100 -bm=default page request total stddev% r/s ms/r html 39.86086 3.2358 2.51 398.61 xml 39.20538 4.8425 2.55 392.05 all requests 79.06625 3.9818 2.53 395.33