Hi,
I'm running qPCR on 50 different samples and if I run them in duplexes I would have to make almost 1000 reactions since the efficiency for the housekeeping gene in each sample should also be calculated.
Does anyone know if it's possible to calculate the qPCR amplification efficiency from only one sample dilution? I have searched the net for possible software programs that are able to calculate this, but haven't found any yet.

This is how we usually do RT-PCR efficiency:

1)make primer/probe aliquots from the stocks
2)create a standard curve (1,10,100,1000 ng) using the aliquots
3)check the slope (-3.23 if I recall correctly)
4)assume that efficiency for all subsequent reactions
5)repeat 1-4 with any new aliquots

The aliquots should maintain that efficiency for at least a year if stored at -20C.

Or maybe I'm not reading your question right...but to create standard curves for each PCR reaction set seems a little crazy.

Isn't it possible to just run 50 reactions each with both your gene of interest and an internal control (18S) together?

You'd calculate the dCT to get the expression relative to the 18S. If you wanted to, you could then run another 50 reactions in which you add varying amounts of cDNA so that the Ct for 18S is the same in each sample and ideally, at an early cycle so that your gene of interest crosses the threshhold before cycle 35 or so.