[Date Prev][Date Next]
Re: Benchmarks at Network World
> Read performance is probably due to searching on
> un-indexed attributes, i.e. no the same ones built during load. Each
> client starts at the same time, starts a thread, and all finish about
> the same time: linear scaling.
If I understand it correctly, this only happens if compiled with threads.
With no threading, each request would block everything and this effect
would not be seen. It would be interesting to know what thread library
they used, if any.