awk     '
BEGIN           {ENG = "BKMGT"
                 for (i=split (ENG, UN, ""); i>0; i--) EXP[UN[i]]=i-1
                }

function NBR(NUMSTR)    {NUMSTR=toupper(NUMSTR)
                         match (NUMSTR, "[" ENG "]")
                         return substr(NUMSTR,1,RSTART-1) * 1024 ^ EXP[substr(NUMSTR, RSTART,1)]
                        }

NR==1           {print $0, "percent"
                 next
                }

                {printf "%s\t %.2f%%\n", $0, 100 * NBR($3) / NBR($2)
                }
' OFS="\t" file