Graeme Rocher (one of the lead Grails guys) has put together a benchmark comparison between a simplistic Rails app and a Grails app. http://docs.codehaus.org/display/GRAILS/Grails+vs+Rails+Benchmark
But there’s some weirdness in the results that people with more Mongrel knowledge than I can help out with. The Rails app was given a cluster of 10 mongrels and the Grails app was limited to a max of 10 threads under tomcat to keep everything even. What’s weird is that Rails performance degraded compared to when it was just run with one mongrel. Obviously there’s something whacked with that and we would really appreciate your insight to make this a totally balanced comparison.
There’s some weirdness on the Java side of things too for any of you with good Tomcat knowledge. For some reason Grails seemed to perform better when Tomcat was limited to 10 threads as opposed to when it ran under the default # of threads.
Feel free to respond here or drop me an e-mail privately and I’ll make sure it gets to the right people. 
