Is there a way to calculate tgammaf(50) on a device using cuda?

I am using stirling's approximation to calculate this, but I am having trouble storing the huge answer I get from Stirling's approximation. Is there any good way to store such a big number on the device?