I’ve submitted two abstracts for the UKOUG 2006 conference in Birmingham.
The first is just the paper that I presented in Dallas in March, Manchester last week and Edinburgh next week. It’ll be a bit like a tour for this one 😉
“How Many Slaves – Parallel Execution and the Magic of 2
Several sources have suggested to me that the Magic of 2, as described by Cary Millsap, comes into play when setting the best degree of parallelism. This presentation discusses tests using various degrees of parallelism against different hardware configurations to investigate what factors should be considered when choosing an appropriate DOP.”
The second is my attempt to address a gaping hole in my last two PX presentations. In both I hinted at trace files and wait events but didn’t really include them as much as I wanted to – mainly due to time constraints. As I’ve been working with people on different PX problems or tests, it has occurred to me that it might be worth addressing the subject properly. I don’t plan on the presentation being super-geeky, but just helpful if you ever have to trace PX jobs yourself. However, it will not include an introduction to PX, so there’ll be no padding if I can help it. (In fact, if Oracle get round to releasing 10gR2 on 32-bit Intel Solaris, I might even try to squeeze some Dtrace in there.) Anyway, here’s the abstract.
“Tracing PX Slaves
How does the anatomy of a parallel query differ from the serial equivalent?
This presentation will take a single parallel query and walk-through the tracing process and resulting trace files. As well as showing how to analyse your own parallel queries using Oracle 10g features, this will illustrate the anatomy of the query and various expected and more unusual wait events.
This is not an introductory presentation, so a basic appreciation of parallel execution will be assumed, although links will be provided to supporting information.”
A couple of final notes. First, I notice that the abstract submission deadline has been extended to this Friday, 5th May, so you still have time to make the decision to present, submit your abstract and keep your fingers crossed. Second, should anyone on the judging panel happen to read this blog, I’d rather do the tracing presentation if I had to pick one 😉 Personally, I just hope I get a chance to present something and meet a few old and new friends while I’m there.