A topic that has popped up several times in the past few weeks is File System Performance. I've seen questions about HP-UX 11i v3 file system performance versus Solaris ZFS. I've seen questions about Logical Volume Manager (LVM) performance. Since many customers run Oracle databases on HP-UX, I've seen questions about the performance of the Concurrent and Direct I/O, especially compared to running Oracle ASM.
HP has a whitepaper [PDF] that discusses the results of performance tests that HP has run with an Oracle single instance database to answer some of these questions. We essentially tested the same setup, a 4 socket BL870c blade system with EVA storage, running the different file system mount options. This represents a realistic, real-world type of implementation, unlike some of the unrealistic set ups for large commercial benchmarks.