So it turns out that I messed up the test a little bit, so the old values that I compared to were worse than they actually should have been. So the new test show like a 2x improvement from the version using HFile which still adds up to 2*5 = 10x, so we are still on track here. Will post some more results when I have the whole system out and running with some better more reliable test. Erik
- First numbers of new Server side implementation. Erik Holstad
- Re: First numbers of new Server side implementation. Ryan Rawson
- Re: First numbers of new Server side implementation. Erik Holstad
