makeblastdb takes a few seconds for a medium size genome. However, see the -subject parameter of blast+ binaries to accomplish this without a db.

It is advisable to create a blast database, because of

• low-complexity filtering and masking (recommended for whole genomes with repeats)
• being able to parse sequence ids (recommended when exported from genbank)
• easy retrieval of sequences by blastdbcmd
• entries can have additional annotation, e.g. individual taxon id (metagenomics)
• you'll always run blast twice

*** BLAST-2-Sequences options
 -subject <File_In>
   Subject sequence(s) to search
    * Incompatible with:  db, gilist, seqidlist, negative_gilist,
   db_soft_mask, db_hard_mask