Model Count for full counting benchmarks

[mardalla]

Hello,

Would it be possible to post a list of all the instances in the counting benchmarks (zip linked in the description) alongside with the exact model count? (for example in a csv file)

Thank you

[msoos]

Hi,

You can re-generate the values from running the now released binary and the example problems that were also released. The binary now is available for both Mac and Linux. In the meanwhile, we can try to look for them.

@smsharma1 do you have the results from these experiments?

Please note that you will get SIGNIFICANTLY better performance if you run Arjun prior to running GANAK or any other model counter, see: https://github.com/meelgroup/arjun Simply run ./arjun myinput.cnf output.cnf and then count output.cnf with GANAK. It will run a LOT faster. So even if we find you the numbers, it may not be indicative of what GANAK can count, since that will be its performance without Arjun.

I hope this helped,

Mate

[mardalla]

Hi Mate,

Thank you very much for your reply. Thanks for suggesting to us Arjun, we were already aware of it and plan to include it in our experiments.

We were interested is the exact model count for the counting benchmarks. We need to use it as a ground truth and unfortunately our computational capabilities are extremely limited and we're not able to extract a model count for all of the instances in those benchmarks.

Thanks again for your help,

Marco