That's quite interesting:
1. SC should not tilt execution time to such a degree - definitely not for a single query running on its own... Sounds fishy.
2. Moving from M4K to M10(Z) should not affect significantly individual query either - unless ASEs hosting the DB are configured differently.
A bunch of questions:
1. Is the old box+ASE still available? Can you run the query there for comparison? If not - what metrics do you have at hands?
2. Are both ASEs configured identically (CFG based comparison)?
3. DId you collect MDAs + plans for the two different executions? I'd start from comparing plans & waits + IO - something in the access path may have changed. If identical - which I doubt - open incident....
Btw, did you move from Solaris 10 to 11? OS level tuning? patches &c...
Please continue to post here - I wonder where this stuff comes from. SC is a funny guy - with a lot of surprises.