| ||||||||||||||||||||
About -- Download -- Statistics | ||||||||||||||||||||
Statistics All benchmarks were done using these versions of the tools:
The instances that we used were from the 2006 QBF-Eval preliminary and fixed set, which is available for download here. We also provide a local copy here. The generated certificate files are available in two versions:
Test Parameters: A time limit of 600 seconds, and a memory limit of 1 GB. |
||||||||||||||||||||
Certificate Generation Time |
||||||||||||||||||||
Certificate Validation Time |
||||||||||||||||||||
Inverted Instances This chart cleary shows that the time needed to solve an inverted instance is almost always considerably longer than the time to solve the non-inverted instance. Actually, a lot of instances that took almost no time to solve normally (<1 sec), timed out when inverted (as seen in the top left of the chart). Result Files: isat.tgz (inverted instances), compstatsQUANTOR-i.csv (Quantor on inverted), compstatsQUANTOR-ni.csv (Quantor on original), compstatsSKIZZO-i.csv (sKizzo on inverted), compstatsSKIZZO-ni.csv (sKizzo on original), compstatsSQUOLEM-i.csv (Squolem on inverted), compstatsSQUOLEM-ni.csv (Squolem on original), |