Closed dilation closed 4 years ago
We should also take a look at the serialized size of models, whether they should be optimized somehow.
Decoding (form ModelRepr
) performance is pathological (probably exponential) for big models involving somewhat long products (probably also sums).
ModelRepr#toModel
. However issues seem to remain: see #88 for reducing the serialized size of models; and see #89 for further improving toModel
performance.
Creating a
Model
from aModelRepr
seems slow. Possible causes:lazy val
s during building the modellazy val
sRef
sMap
s instead ofHCons
?