finished updated 2026-11-07 22:43:42 UTC

Summary

CUDA current
1582/1582
1405 solved
CPU current
1582/1582
1418 solved
setdeviceproblems Ipoptmastercurrent
solvedSGM (s)SSGMsolvedSGM (s)SSGMsolvedSGM (s)SSGM
ACOPFCUDA132/13212319.5711245.650.5291245.860.537
COPSCUDA65/656121.161653.170.423653.040.419
CUTEstCUDA1338/133812057.56111807.781.01
MittelmannCUDA47/473989.6713631.360.415
ACOPFCPU132/13212524.27112619.660.866
COPSCPU65/656423.2916417.610.829
CUTEstCPU1338/1338122311.891118516.581.21
MittelmannCPU47/4743126.14143109.430.877

Solvers

solverdevicerevisionbenchmark
IpoptCPU3.14.19–
masterCUDA9bbe04800.1.0
currentCPUd9940730 to 315765320.1.0
currentCUDAac03fdde to 315765320.1.0

CUDA: shin-compute-005: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz (40 cores, 80 threads), 123.3 GB of memory; Quadro RTX 6000, 23.5 GB, driver 13.5.0, CUDA runtime 13.4.0

CPU: shin-compute-005: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz (40 cores, 80 threads), 123.3 GB of memory

Benchmark 0.1.0: the version fixes the instances, the options, Ipopt and HSL.

Base: `9bbe0480`, on Quadro RTX 6000 (acopf, cops)

SGM: shifted geometric mean of the solve time, with a shift of 10 s and a failure counting as the time limit of a solve (CPU 7200 s, GPU 600 s, for every solver on those pages), over the problems that all solvers shown have run. SSGM: the SGM scaled so that Ipopt, or the best solver where Ipopt has not run, is 1.