On 8/18/10 12:24 PM, Samuel Gendler wrote:
> With barriers off, I saw a transaction rate of about 1200. With
> barriers on, it was closer to 1050. The test had a concurrency of 40
> in both cases.
I discovered there is roughly 10-20% "noise" in pgbench results after running the exact same test over a 24-hour period on a machine with no other activity. Be sure you run your tests enough times to get good statistics unless you're looking at much larger differences.
Craig