CUDACyclone: New GPU Mining Tool with Impressive Speed

Hey folks! Just wanted to share the new CUDACyclone tool. It's super lightweight and hits 1.3 Gkeys/s on the RTX4060. The best part? It uses very little VRAM on rented GPUs. Works even if Vanity or Keyhunt don’t run. Plus, it's a neat learning resource, only 7 small files. Check it out! Here's a quick test run for the RTX4090:./CUDACyclone range 400000000000000000:7fffffffffffffffff address 1PWo3JeB9jrGwfHDNpdGK54CRas/fsVzXU grid 512,512 Just some info about the GPU...

Not trying to be rude, but what's up with JLP's copyright and license? Do you even know what GPL means?

KtimesG, good to see you! Yeah, honestly not too familiar with licenses. I did add the link to JLP from the beginning in my repo, along with FixedPaul. These guys are great at coding. If you need more proof, I can add more info about them in my repo. Totally not a thief here! Just so you know, I only used a part of CUDAMath from JLP.

Can someone with skills check the code? I’ve heard many tools related to Bitcoin puzzles have backdoors. You might think you found a Private Key, but it could be sending data to the tool owner instead. I see a lot of detail in the GitHub, but it could use more context for trust. What makes your tool better than KeyHunt or BitCrack?

You can check it yourself or even ask AI to analyze it for backdoors. VanitysearchA-bitcrack from GixedPaul is the fastest I saw, hitting 5.15 Gkeys/s on RTX4090. I got up to 6.2 Gkeys/s. But I got an issue too! GPU frequency goes up, but the speed drops after an hour (like from 6.2 to 4.9). It’s a compute-bound issue with power throttling. Classic PL/TDP situation SM-clock increases, Mkeys/s decrease while mem-clock stays the same. I’ll have a fix after a few days, just need to stabilize GPU frequency.

Hey FrozenThroneGuy, Awesome work, man! Thanks for sharing. I ported it to a Windows exe using Visual Studio with Chat GPT. Couldn’t keep it for Linux though, so it’s just a Windows project. I tested it on a Frankenstein RTX 3060M. Here’s what I got: C:\CUDAWindowsPort_VS2022\x64\Release>CUDAWindowsPort range 1000:22382FFFFF address 1HBtApAFA9B2YZw3G2YKSMCtb3dVnjuNe2 grid 512,256 Device info for the GPU...