RANDOM mode does not enforce any sequence or minimal number of keys to generate before checking. Each key is generated and checked independently.

The program does not enforce any specific sequence or pattern in the keys it generates. It relies on the randomness of the PRNG to explore the key space uniformly.

The program processes keys in batches of 512 (2 × 256) to optimize performance using AVX2 instructions for hashing.

Progress and elapsed time are not calculated in RANDOM mode.

The speed of 4.6 Mkeys/s per core is achievable with AVX2 optimizations, but it depends on the hardware.


7985WX has a base clock frequency of 3.2 GHz and a maximum boost clock frequency of up to 5.1 GHz for a single core, depending on workload and thermal conditions.

Yes, I am using affine coordinates for batched addition, as it is more efficient for certain operations.


I am going fishing today.

That's weird. I have over 17 MK/s per core on a Laptop CPU: i9 13900H and it's only running at 3 GHz.

This with a batch size of 512 constant elements, adding a single point (middle) left and right in a range of 1025. Higher batch sizes have diminishing returns, below measurable.

This is futile though, since the hashing is the bottleneck, it will never matter if the EC throughput increases.

Let's assume we can hash at 6 MH/s per core.

With 12 MK/s EC throughput/core, we end up with (12 * 6) (12 + 6) = 4.0 M op/s
With 24 MK/s EC throughput/core, we end up with (24 * 6) (24 + 6) = 4.8 M op/s
With 100 MK/s EC throughput/core, we end up with (100 * 6) (100 + 6) = 5.66 M op/s

etc...

I am currently using and testing AVX-256(for these headache sufferers), and it is slower than AVX-512 by half. The AVX-256 workload here can cause the CPU to downclock significantly, which might negate the performance gains from vectorization. The CPU is throttling as a result.



Your brain is like that one friend who loves drama but can’t feel it themselves. It’s sitting there going, ‘Oh no, the nerves and muscles around me are freaking out! Better make you feel like your head’s exploding!’ But your brain? Totally chill, no pain receptors, just vibes. So when your head hurts, blame the drama queens around your brain, not the brain itself.  

Fixed.
Added modulo operation and `minKey` addition to constrain the random key to the correct range.

https://github.com/NoMachine1/Cyclone/commit/3627436f543e5e3919b2c9ce6dde320474bf9c79

Great, thanks!

Look your PM

Howbout cyclone on cuda 🤔

Stride will not work well over 50 bits because the type used for batches is uint32_t . A large stride might cause the batch size to exceed the limit, leading to incomplete or incorrect processing, inaccurate progress calculations, or even crashing the entire program. This is why I prefer not to modify other people’s scripts.


I work with one hand and one eye, so mistakes are inevitable. Here's the fix:
https://github.com/NoMachine1/Cyclone/tree/main

collecting some entropy..

https://www.daskeyboard.com/blog/wp-content/uploads/giphy.gif

AMD Ryzen 9 7900 12 x 3,70 GHz