In smaller range with mulitple prefixes, I actually drop the x grid size.
I match it up with the actual x size. So for 3070 it is 46 so I put in a -g 46. But you can tweak to your liking. Have you tried one prefix/address?
If it could use full address and search for RIPE, would be faster.
in full address
C:\VanBitCracken\VanBitCrackenS1.exe -gpu --keyspace 8000000000000000:ffffffffffffffff 16jY7qLJnxb7CHZyqBP8qca9d51gAjyXQN
VanBitCrackenS v1.0
Keyspace start=8000000000000000
Keyspace end=FFFFFFFFFFFFFFFF
Difficulty: 1461501637330902918203684832716283019655932542976
Search: 16jY7qLJnxb7CHZyqBP8qca9d51gAjyXQN [Compressed]
Started at Sat Apr 3 02:26:04 2021
CPU threads used: 5
GPU: GPU #0 GeForce RTX 3060 Ti (38x128 cores) Grid(304x512)
794.487 MK/s (GPU 775.923 MK/s) (2^32.19) [00:00:06 Elapsed Time][0]GPUEngine: Launch: misaligned address
Finish at Sat Apr 3 02:26:13 2021
Right, as I said, must use string vs full address. I usally chop 7 to 8 characters off.
That's what I don't get. Everyone trying to tweak the other program which won't even run with updated drivers, with any 20xx or 30xx cards, and try to fix the misaligned address issue. Where as VS/VBC only has the issue of using full address. Maybe everyone loves the "stride" version as you stated
Anywho, hopefully someone figures it all out or Jean Luc comes back to save the day!
Can you release 512, 256 and 128 also
I ran those for comparisons but the 512 always got through a range faster. I even recompiled newer version where you could manually enter -g x,y and it didn't even perform as well.