Closed izzayacine closed 1 year ago
Hi,
Currently, unfortunately, this is not supported. Would this be useful for your use-case? Is there a particular use-case you are looking for?
Thanks,
Mate
My program instruments a linear number of call to approxmc (on the same cnf formula + assumptions, so different sub-formulas) and I want to fix a time/memory limit for each call, as some calls consume too much resources. Therefore, in case of an expensive call, the procedure should skip it.
Yacine
Hi,
Let's talk about this once I'm back in Singapore, in about a week. It's just a lot of work and I'm not 10% sure it's needed. Let's discuss in-person and then we can see if it's indeed needed.
Cheers,
Mate
The issue is that when I run approxmc with python interface it stops the process (killed -9) however the binary file (c++ compiled) returns a results
Hi,
Oh, that's a bug actually. Can you please send me how to reproduce?
Thanks,
Mate
On Wed, 3 May 2023 at 18:50, Yacine Izza @.***> wrote:
The issue is that when I run approxmc with python interface it stops the process (killed -9) however the binary file (c++ compiled) returns a results
— Reply to this email directly, view it on GitHub https://github.com/meelgroup/approxmc/issues/35#issuecomment-1533973481, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAKF4OO2TJVA4B2KHPYXQS3XEMDN3ANCNFSM6AAAAAAXTYWAQM . You are receiving this because you commented.Message ID: @.***>
Okay, I sent you an email.
Yacine
The solution is to use the projection
parameter to count()
in the python module. This is equivalent to the c ind
parameter in the CNF. Without a projection set, ApproxMC counts significantly slower.
I am now closing this issue, since it's not a bug in ApproxMC.
Mate
I use the param projection when calling the function count() and I get an error. Have you tried to replicate with the formula that I've sent? and do you get an error killed -9 or it returns the solution?
I think that this occurs only when I call several times approxmc, and then the memory gets stuck. I try to delete the instance of Counter class but it doesn't solve the problem.
Can we fix the memory and time limit of the Counter using the python interface?