Ahh, I reread your first post and there is only one stick of RAM. That would at least explain part of the performance issue. Going to 4 GB when running in dual channel mode should help.

Edit...Also, running quite a few hadam3p models, you are probably running out of memory, so going to 4 GB should also help in that regard.

If you are running eight threads on an i7 then is it reasonable to expect each thread to be faster than each core on a Q9550? Hyperthreads are not cores.

You should get more credits in total from the i7 machine than from the Q9550 machine because each i7 core is faster than each Q9550 core and running two hyperthreads on each i7 core gives a further multiple - it used to be about 1.2 (i.e. much less than 2.0).

Currently only hadsm3 models are available. If you have other model types checked in your climateprediction.net preferences, then perhaps that is why you are not receiving any more work?

I'm not trying to be a bossy boots about this, it's just that it makes it hard to tell just what people are reporting problems with.

Is anyone having a problem with trickles missing?
Or is it just no credits recently? (Due to the trickles-to-credits program not running, which gets re-started before long, and results in a rush of credit.)

e.g. I've just uploaded 3 trickles each for 8 models, and all of these trickles appear on the pages for their models.