Skip to content

Instantly share code, notes, and snippets.

View theStack's full-sized avatar
💭
BUIDL

Sebastian Falbesoner theStack

💭
BUIDL
View GitHub Profile
@theStack
theStack / secp_pr1767_bench_results.txt
Created November 7, 2025 20:30
Benchmark results for secp PR #1767 on my arm64 notebook (Lenovo Thinkpad T14s Gen 6, with a Qualcomm Snapdragon X Elite CPU)
$ cc --version | head -1
cc (Ubuntu 14.2.0-19ubuntu2) 14.2.0
$ lscpu | head -3
Architecture: aarch64
CPU op-mode(s): 64-bit
Byte Order: Little Endian
$ uname -a
Linux thestack-ThinkPad-T14s-Gen-6 6.17.0-8-qcom-x1e #8-Ubuntu SMP PREEMPT_DYNAMIC Sun Aug 31 21:03:54 UTC 2025 aarch64 aarch64 aarch64 GNU/Linux
===== Baseline =====
@theStack
theStack / bip352_scanning_approaches.txt
Last active December 9, 2025 07:26
Different approaches for scanning in Silent Payments (BIP-352)
Different approaches for scanning in Silent Payments (BIP-352)
=============================================================================================
===== "calculate label candidates for each tx output, look them up in the labels cache" =====
=============================================================================================
The current approach for scanning in secp #1765 is as follows, following BIP-352:
For each k value, derive the unlabeled output candidate P_k, then loop over the
transaction x-only outputs linearly to check whether there is either a unlabeled or a
labeled match. For the latter, two label candidates are calculated for each transaction