The latest update to the EPFL benchmark exposes some error in our Picotrav benchmark. In particular, epfl/best_results/size/int2float_size_2025.blif results in a gate differs error.
Output gate ['M[3]'/'y3'] differs!
Presumably, this is an error on our end. This new input is, compared to the previous inputs, different in some way with respect to what features of the BLIF format is used; there is some edge case we probably did not implement correctly.
The latest update to the EPFL benchmark exposes some error in our Picotrav benchmark. In particular,
epfl/best_results/size/int2float_size_2025.blifresults in a gate differs error.Presumably, this is an error on our end. This new input is, compared to the previous inputs, different in some way with respect to what features of the BLIF format is used; there is some edge case we probably did not implement correctly.