Thanks, I have checked Shuffle-LAGAN, seems to be a good program and I will test it later. My only concern is that was designed to do glocal alignments of long DNA sequences, while I have short-medium protein sequences (20-500 aa long).
I am looking for a program that would be able to perform semi-global (global in query, local in reference sequence) protein alignments fast and efficiently. My query sequences are short (20-30 aa), while reference sequences 10-20 times longer. So far I have been able to find GLSEARCH (from FASTA package) and align, but they seem to be somewhat slow for my task. I would really appreciate it if someone could suggest another alternative.

3 answers
At least one software that strongly uses the "glocal" terminology is LAGAN, specifically Shuffle-LAGAN
e.g. this website http://genome.lbl.gov/cgi-bin/VistaInput?align_pgm=slagan&num_seqs=2 uses the phrase
Shuffle-LAGAN Glocal pair-wise alignment of finished sequences (detects rearrangements)
Parasail (https://github.com/jeffdaily/parasail)
Thanks, I will check it.
Nucmer, part of the Mummer package (http://mummer.sourceforge.net/).
Please correct me if I am wrong, but I had an impression that nucmer performs global alignments of long sequences, while I need semiglobal (glocal) alignments of short to moderately long sequences. I updated my original post to reflect the difference between different methods of alignment.
Log in to answer this question.