Wow, that cadence is really low! When I hit a mountain that forces me into the 50rpm range, I start debating just hoofing it for a while instead of pedaling that slowly. I did discover that my maximum sustainable mountain climbing effort is 70rmps at 180bpm HR. Below that cadence, I start losing it, and I can't keep the heart rate under control. That's assuming that I've run out of gears by then, given the grade. Higher cadences on an easier grade is my preference. Plus, I think it's more effective. For example, a friend and I were tackling some mountains in WV recently, and we both have the same lowest gear. She got a head start on one of them, because she could descend a bit faster off the previous mountain. This was a relatively easy climb (in terms of grade, not length) compared to other stuff we'd hit. I zoomed past her easily less than half way up, because I was spinning about 85rpms in my easiest gear. I was baffled at why I was passing her so easily, and I look down and see she's 2-3 cogs harder than me, grinding away with a low cadence and suffering. For the really long climbs, it's just hard to use brute strength to get up them. Ideally, if we were to all practice on that terrain regularly, we'd both be able to spin fast in that harder gear. In my experience, it is harder to improve climbing endurance trying to do overgeared strength climbs than it is to work on getting your heart rate and muscles to handle spinning. I started out wanting to power up hills, because my body likes to work that way, but inevitably, I'd blow up, and I'd never get faster. It hurts like mad to work on spinning extra fast uphill, but in the long run, I think it balances out that flatland strength and power in helping the non-climber types like me learn to find some climbing legs.

That said, I agree with AR that the best way to improve on those types of climbs is to find them and climb them over and over. It is too hard to simulate it on a bunch of short hills or on the trainer, because you need to work on climbing and strength, and, at first, do the two pretty much separately.