(up to around 44 bits), random is faster, in my tests.
Up to around 44 bits database is faster. The point is that why do you care about the speed in low bit problems? 2^64 and up is more interesting. My mod is based on bitcrack and yes it is scanning bruteforcing all of the solutions. That might be a littlebit stupid, but stupid solutions can still solve problems with enough hashing power. My fork has a random mode. Run with -r --random.
See, here you go again. You posted results of 34 bit and a 39 bit. I was merely following your lead and comparing programs.
So you have a database that has 17,592,186,044,416 keys in it? Why? And you say it is faster but how long does it take to compute and store 17,592,186,044,416 private keys?
And I don't care about speed at all; I care about results...in time.