Participants Info
| Name | Ting Li, Yao Sun |
|---|---|
| Institute | State Key Laboratory of Cyberspace Security Defense, Institute of Information Engineering, CAS |
Submission Details
| Date | 2026/09/17 |
|---|---|
| Type | V |
| Number of variables (n) | 31 |
| Number of equations (m) | 21 |
| Seed (0,1,2,3,4) | 0 |
| Algorithm | Improved F4 via smaller matrices |
| Hardware | 2 x RTX 5090 (32 GB each), 16 CPU threads (Intel 6530, 2.1 GHz), 256 GB RAM |
| Running Time | 722,188 seconds (8 days 8 hours 36 minutes) |
| Answer v=[v1,…,vn] in Fn | [4f, a8, 89, 4e, e5, e5, 24, 97, 62, 56, 04, 96, 02, 76, 26, 36, f0, 76, 4b, ce, f1, ec, 95, a6, 35, 0f, c6, 85, 9e, 09, b0] |
| Notes | We randomly assigned values to 11 variables, reducing the original system of 31 variables and 21 equations to subsystems of 20 variables and 21 equations. Each instance was processed using 16 CPU threads and two NVIDIA GeForce RTX 5090 GPUs. A verified solution was found after completing 80 instances. The average wall-clock time per instance was approximately 9,027 seconds. We used a modified F4 algorithm that constructs more compact matrices, thereby reducing the computational cost of solving each subsystem. We will describe the algorithm in detail in a future paper. |