For the build errors are you building from this repo - https://github.com/hibernating-rhinos/rhino-etl ?
Nathan Palmer On Wed, Feb 23, 2011 at 2:07 AM, F.D.Castel <fdcas...@gmail.com> wrote: > I was playing with Row class implementation from Rhino.Etl and found a > nice surprise. > > I did a new test "Test_Row_performance" on RowTests.cs. This is the > result for the original implementation: > > > * Row-original.cs / QuackingDictionary-original.cs > Normal Invocation took 2273 milliseconds > Normal Invocation took 2285 milliseconds > Normal Invocation took 2269 milliseconds > > > Then, I changed QuackingDictionary to inherit from DynamicObject and > implemented the necessary methods (similar to IQuackFu). The > performance for normal invocation suffered a little (+~100ms). And the > new "dynamic" invocation is (as expected) terrible :) > > > * Row-dynamic-hashtable.cs / QuackingDictionary-dynamic-hashtable.cs > Normal Invocation took 2332 milliseconds > Dynamic Invocation took 3335 milliseconds > Normal Invocation took 2312 milliseconds > Dynamic Invocation took 3294 milliseconds > Normal Invocation took 2313 milliseconds > Dynamic Invocation took 3301 milliseconds > > > Finally, I also changed the base implementation from a HashTable to a > Dictionary<string,object>. This time, even keeping the dynamic > support, the results were interesting: > > > * Row-dynamic-dictionary.cs / QuackingDictionary-dynamic-dictionary.cs > Normal Invocation took 1788 milliseconds > Dynamic Invocation took 2760 milliseconds > Normal Invocation took 1779 milliseconds > Dynamic Invocation took 2771 milliseconds > Normal Invocation took 1784 milliseconds > Dynamic Invocation took 2778 milliseconds > > > A good ~20% improvement over the original. And the dynamic, while yet > slower than the static, is now considerably better. Just 20% slower > than the original HashTable ;) > (But, of course, yet a hard price to pay if we're talking about > millions of rows...) > The implementations and the tests are in the attached .zip file. I > would love to hear any opinions about the subject. > > Best Regards, > Fabio. > > > > P.s.: BTW: The last commit on github isn't building on VS. There are > some .Block -> .Body changes needed on Rhino.Etl.Dsl > > -- > You received this message because you are subscribed to the Google Groups > "Rhino Tools Dev" group. > To post to this group, send email to rhino-tools-dev@googlegroups.com. > To unsubscribe from this group, send email to > rhino-tools-dev+unsubscr...@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/rhino-tools-dev?hl=en. > > -- You received this message because you are subscribed to the Google Groups "Rhino Tools Dev" group. To post to this group, send email to rhino-tools-dev@googlegroups.com. To unsubscribe from this group, send email to rhino-tools-dev+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/rhino-tools-dev?hl=en.
