Fortunately, I have some files saved spanning the problem/omission, so here's some hard comparative numbers, using the same input file:

./qsafe 64 < n1024q0 > n1024s0
# Mon Jun 30 17:42:43 2003 Final 64 Miller-Rabin trials (0 generator)
# Mon Jun 30 18:17:27 2003 Found 7 safe primes of 12816 candidates in 2084 seconds

./qsafe 64 < n1024q0 > n1024s0x
# Sun Oct  5 23:56:35 2003 Final 64 Miller-Rabin trials (0 generator)
# Mon Oct  6 02:40:39 2003 Found 7 safe primes of 12816 candidates in 9844 seconds

./qsafe 64 < n1024q0 > n1024s0xx
# Thu Nov  6 12:57:26 2003 Final 64 Miller-Rabin trials (0 generator)
# Thu Nov  6 13:32:28 2003 Found 7 safe primes of 12816 candidates in 2102 seconds

As you can see, the problem has been fixed (for i386) in 
  http://mail-index.netbsd.org/source-changes/2003/11/04/0037.html
-- 
William Allen Simpson
    Key fingerprint =  17 40 5E 67 15 6F 31 26  DD 0D B9 9B 6A 15 2C 32