shrecshrec

4 ms·

1. 1.▲
   by shrecshrec 7mo ago

   i'm working on this two projects: https://github.com/shrec/Exorcist and https://github.com/shrec/UltrafastSecp256k1
2. 2.▲
   We Built UltrafastSecp256k1 Up to 51% Faster ECC Across x86,ARM64,and RISC-V

   5 points by shrecshrec 8mo ago|0 comments
3. 3.▲
   UltrafastSecp256k1 – 4.88M ECDSA signs/s on GPU zero dependencies 12 platforms(github.com)

   2 points by shrecshrec 8mo ago|1 comments
4. 4.▲
   by shrecshrec 8mo ago

   UltrafastSecp256k1 is an open-source C++20 secp256k1 library built for throughput. No dependencies — no Boost, no OpenSSL. Just a conforming C++20 compiler. GPU benchmarks (RTX 5060 Ti, kernel-level): • ECDSA Sign: 4.88 M/s (204.8
5. 5.▲
   UltrafastSecp256k1 Zero-dep C++20 secp256k1 with ASM,CUDA, 27 coins,MuSig2,FROST(github.com)

   2 points by shrecshrec 8mo ago|1 comments
6. 6.▲
   by shrecshrec 8mo ago

   I've built a comprehensive secp256k1 library in C++20 with zero external dependencies. Everything — SHA-256, SHA-512, Keccak-256, RIPEMD-160, Base58, Bech32 — is implemented from scratch. Key highlights: Assembly backends for x64 (MASM
7. 7.▲
   by shrecshrec 8mo ago

   I will be glad to hire any suggestions from everyone abut future improvements and ideas.
8. 8.▲
   Building a Zero-Dependency secp256k1 CUDA Engine from Scratch (2.5B ops/SEC)(github.com)

   2 points by shrecshrec 8mo ago|2 comments
9. 9.▲
   by shrecshrec 8mo ago

   I implemented a full secp256k1 engine from scratch in C++ and CUDA with zero external dependencies (no GMP, no OpenSSL). The goal was to explore performance limits of: Jacobian mixed-add Batch inversion using Montgomery’s trick Large-scale