Quote from: fmg75 on September 21, 2025, 10:12:41 PM

Quote from: pp-duster on September 21, 2025, 07:10:50 PM

Quote from: fmg75 on September 21, 2025, 01:26:46 PM

Quote from: pp-duster on September 21, 2025, 05:12:14 AM

Quote from: FrozenThroneGuy on September 20, 2025, 09:14:54 AM

Quote from: fmg75 on September 18, 2025, 08:39:47 PM

An update to CUDACyclone that generates non-repeating, persistent random subranges.
Ideal for collaborative search.

./CUDACyclone \
--range 400000000000000000:7fffffffffffffffff \
--address 1PWo3JeB9jrGwfHDNpdGK54CRas7fsVzXU \
--random-subranges --num-subranges 10000000 --grid 256,8

Hello man, update your software. Previous version of a Cyclone has a bug. I did a mistake with a first batch of the thread, not right computing center point of the first batch, and than the last batch finish not in the end of the range. Last half batch of points from the end of the each thread doesnt computing!

Hi, what was exactly the bug in the software? How would I rescan the already covered ranges?

Suppose my command with older version was

Code:

./CUDACyclone --range 40003da10000000000:40003da1ffffffffff --address 1PWo3JeB9jrGwfHDNpdGK54CRas7fsVzXU --grid 512,512

Do I need to run this range once again with

Code:

./CUDACyclone --range 40003da10000000000:40003da100000000ff --address 1PWo3JeB9jrGwfHDNpdGK54CRas7fsVzXU --grid 512,512

and

Code:

./CUDACyclone --range 40003da1ffffffff00:40003da1ffffffffff --address 1PWo3JeB9jrGwfHDNpdGK54CRas7fsVzXU --grid 512,512

to recheck for the missing keys?

I want to clarify that my post refers to a modified version of the original CUDACyclone project, where I added the features I mentioned.
I don't know what bugs @FrozenThroneGuy is referring to.

Idk, I have addressed my message to the owner of https://github.com/Dookoo2/CUDACyclone who admitted in code that there was an issue with missing keys. I'd want to know the workaround strategies to not fully rescan the ranges which were previously scanned with the bugged software.

The updated code fixes issues with lost keys at the ends of the ranges.
But I'm not sure if that's the only problem.
I don't see a reliable way to 100% guarantee that there are no lost keys in other parts of the range.

The great challenge of these brute-force implementations is probabilistically minimizing the chances of key "losses." In such large search spaces, this is very difficult. You can never be 100% sure that you have verified all the keys. There is much work to improve these algorithms.