On Wed, 2005-11-23 at 18:20 +0000, Podz wrote: > On a regular blog, of how much value is this cache behaviour ? At what > level of hits does it become something postively useful as opposed to > something bolted on ?
Running Apache Bench against my blogs showed a modest improvement of 1 - 3% for most runs. Larger numbers of requests and increasing concurrency showed the most benefit. When I really hammered the server, load average was less with the cache turned on and the longest response time was much shorter. Please run benchmarks and see how it handles. If the persistent cache doesn't buy us much, we can consider turning persistence off by default. (By persistence I mean saving the objects to disk. We always cache in memory. Even without saving to disk, there is a cache and it is hit hundreds of times for each page load.) Even if the persistent cache doesn't dramatically increase performance for most, the code and API cleanup that resulted from adding it are worthwhile. It's not much more code than before, it's all in one place, it degrades nicely when persistence is disabled, and it's completely pluggable. Regardless, users shouldn't care about the cache. Ryan _______________________________________________ wp-testers mailing list [email protected] http://lists.automattic.com/mailman/listinfo/wp-testers
