make sure your program is working correctly
In terms of testing and correctness, it passes every test I can think of. I'm not sure what else to throw at it. There are likely bugs in the program that I haven't been able to find yet.
Ill rent a cloud 1080Ti real quick and let you know
Edit : Only 1080Ti I could find on vast clore are power limited so I cant run a real benchmark for you. I have a 1080Ti somewhere at my house, but swaping my 5090 is not worth the hassle
Thank you for the quick reply and for trying to check the benchmark. Yes, swapping the 5090 for a 1080 Ti, and possibly dealing with driver changes, isnt very appealing, especially for some random person on the internet like me. This is the only GPU I have available, so Im limited to 2017-era hardware. Ill try to squeeze a bit more performance out of my 1080 Ti, though I'm not sure how much is possible at this point.
For a quick comparison between hashing and non-hashing (hashing EC coordinates vs. computing only EC coordinates), the difference is substantial. Using the lowest values: 760M hashes/s vs. 1.6B EC coordinates/s (both are measurements of comparisons per second). This means operations without hashing are roughly 2.1 times faster than those with hashing. On my 1080 Ti. Improving either part would result in more processed keys and addresses, but Im at a point where even a 1% gain is a massive win for me, because Im stuck on what to possibly improve. I guess I have to try harder and learn more.
Overall, I was just wondering if Im on the right track. Based on publicly available benchmarks, I seem to be, but not everything is public, which is why I asked here. I suppose the next step is finally upgrading my PC or, as I mentioned, trying harder to improve the code.
