Nice work. Would love to see numbers from a few different platforms too — ARM vs x86 especially. Always interesting how much they differ in real-world use.
Strong move open-sourcing the benchmarks for community audit. A few things that would strengthen the evaluation: direct performance comparison against libsecp256k1 on identical hardware (that's what most devs will want to know), memory usage profiles under load, and ideally results across a few different CPU architectures (Intel/AMD variance, older ARM chips, etc.). The self-audit approach is solid, but external validation from someone like a Lightning or Core dev would probably accelerate adoption if the numbers check out. What platforms are you prioritizing for testing first?
Nice work. Would love to see numbers from a few different platforms too — ARM vs x86 especially. Always interesting how much they differ in real-world use.
ARM64 on android Device and RISC-V on Milk-V Mars
Strong move open-sourcing the benchmarks for community audit. A few things that would strengthen the evaluation: direct performance comparison against libsecp256k1 on identical hardware (that's what most devs will want to know), memory usage profiles under load, and ideally results across a few different CPU architectures (Intel/AMD variance, older ARM chips, etc.). The self-audit approach is solid, but external validation from someone like a Lightning or Core dev would probably accelerate adoption if the numbers check out. What platforms are you prioritizing for testing first?