The M5000 with SPARCVII+ is two generations newer then the p570. So it should be "faster" in general vs a p5+ p570.
If you are using LPARS a p570 is quite a different machine then a M5000. To get the same number of virtual instances on the M5000 you'll need to use containers. Which may or may not be better for you the the p570 laprs depending on your use case.
Depending on the requirements I'd probably split the M5000 into its two system domains, one for prod the other for test and then break those system domains down into containers to host multiple instances of Solaris.
Thanks! We're actually planning on doing that (re: splitting it into 2 domains as you suggested).
Is it more comparable to the p6 chip? If so, are there benchmarks for Oracle performance? I was hoping to compare apples to apples to see what we need to buy. We have a total of 16 dual core p5 processors for production, and were planning on going 16 cores for the M5000.
I think no matter what you're comparing an apple and an orange, the machines are just too different to design a one-to-one comparision. Even if you attempt o compare to the more modern power platforms (p6 570 or p780).
I'm rather confident that if your workload runs well on 16 p5 cores it will run on 16 SPARCVII+ cores similiar or better. More likely disk and amount of memory deciated to the database will have a greater affect then CPU speeds.
If your looking to do a more complex TCO comparision of new power 7 hardware and new sparc hardware you're asking in the wrong place, there's just to many variables to cover in a forum discussion. Software and personnal costs will probably dominate the discussion.