It took me 45 hrs to map a 24G single-end RNA-seq data to human hg19 with the gtf file from GENECODE!
I run tophat on a cluster with 1 node, 8 processors with 3gb ram for each processor (the max free resource as a regular user). I have six such files...it will take me 12 days to finish the mapping....wish I had better access to the computing resource...
a quick google confirmed that it does take that long...
http://seqanswers.com/forums/showthread.php?t=5148
"It routinely takes ~40 hours for me (~45 million reads, with - p 4 to use four threads and 16G RAM). It will delete the huge temp files it generates, so the disk usage may not be that much of any issue."