search

lsils / date2020_experiments

6 stars 2 forks source link

Segmentation Fault in optimized cases #1

Open devildonkey1392 opened 3 years ago

devildonkey1392 commented 3 years ago

Hi,

The commit 510579a of mockturtle seems to be deleted.

Therefore, I checkout the mockturtle to a1f0f90 which is also from pull request #231.

However, when I try to run the experiments, it encounters segmentation fault while doing resubstitution in some cases.

For those can't further optimize, the program can finish successfully.

How to fix this problem ?

Thanks a lot!

hriener commented 3 years ago

Could you be more specific?

It takes quite some time to run the experiments on my computer. Which one segfaults?

devildonkey1392 commented 3 years ago

Hi,

Thanks for your response.

I download the source code of version(510579a) and build the code again.

It can work right now!

As for version(a1f0f90), I try the cavlc and div. Both of them get segmentation fault while doing resubstitution.

hriener commented 3 years ago

I download the source code of version(510579a) and build the code again.

It can work right now!

Means that the issue has been resolved?

hriener commented 3 years ago

As a side note, I can reproduce the segfault with benchmark cavlc_untilsat on commit a1f0f90.

You could try this code on the branch master. I recently fixed some pointer issues in the event API. This could be the problem that you are hitting in the old implementation. Note that on the master branch, cut_rewriting has to be replaced with cut_rewriting_with_compatibility_graph. Also, note that a couple of changes have been made to the resubstitution algorithm. I'm not sure if there is any impact on these experiments.