TPC-H Benchmarking
I've got a new 11gR2 environment set up, and I wanted to do some benchmarking. I have never had the time to do it in the past, so I want to do this right.
I set up the TPC-H schema using Hammerora. Hammerora is working perfectly, I can set up virtual users and they will submit the built in 22 query set on demand.
My question: How is queries per hour (qph-h) generally calculated for the 'real' benchmarks? Do you just sum up the number of queries fired off, and the Hammerora reported response time? Do they calculate queries per hour off something on the server? I don't went to start a tracing session, as that would impact the benchmarking.
I suppose all of this is academic anyway, if I am just consistent how I conduct the test across environments and changes to them. I was wondering what I would need to do to get as close to the industry accepted procedure as possible, so my numbers are somewhat meaningful outside my environment.