https://github.com/wurstbroteater/ddueruemma
Pre Release Version for implemeting static variable orderings, evaluated in my master thesis
https://github.com/wurstbroteater/ddueruemma
Last synced: 4 months ago
JSON representation
Pre Release Version for implemeting static variable orderings, evaluated in my master thesis
- Host: GitHub
- URL: https://github.com/wurstbroteater/ddueruemma
- Owner: wurstbroteater
- Created: 2022-09-05T06:51:28.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2022-11-30T12:13:42.000Z (over 2 years ago)
- Last Synced: 2025-01-06T21:46:40.630Z (5 months ago)
- Language: Python
- Homepage:
- Size: 36 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# ddueruem
Pre Release Version for implementing static variable orderings evaluated in Eric's Master Thesis.
### Requirements
* Python 3.8+
* `make`
* `cmake`
* `glibc`
* `libgmp-dev`
### Usefull commands
That I should have known before using try-and-error to find them ...
#### Initialization and Installation
```bash
./ddueruem.py --install smarch
```
```bash
./ddueruem.py --install cudd
```
### New SVOs
- pre_cl_size
- pre_cl_min_span
- fm_traversal_bf
- fm_traversal_df
### New Functionalities
- FORCE without timelimit (if timelimit is -1)
- If input file is xml, the corresponding file types`dimacs`, `uvl` and `sxfm` can be created automatically
- `.orders` and `.order` file now have own order_parser
- Build BDD from order file (for this args should not include any --svo ... )
- `initFeatureModels.sh` creates a folder `evalutation` in the workspace, containg all the feature models used in the SVO evaluation. Dir structure for every model is `evaluation//.xml`
Added new SVOs and integrated them such that they can be used without any overhead for the user.
#### Discussion
We have to discuss the following issues:
- `Pre_cl_family` and `fm_traversal` do not use the class `FM`. Refactoring to use it could be cumbersome and therefore it is discussable if this is really necessary.
- If we agree to the need of using the class `FM`, we should also consider to create a Class for the clusters used in Pre_CL
- `ddueruem.py` needs a slight rework because currently there are checks for file existens and if a file exists, e.g., building BDDs wont be executed. ~~Furthermore, there is the list `evals` which collects all filepaths to the evaluation models, filters them and sortes them according to file sizes. We could remove this or maybe introduce a flag like --evaluation.~~ New ideas are always welcome!
- ~~Extend force such that it can handle xml files (create dimcas file if input is xml)~~ DONE
#### Usage Examples
Fresh start of evaluation for pre_cl_min_span
```bash
./init_featureModels.sh
./ddueruem.py --evaluation --svo pre_cl_min_span n:1
```
A certain model
```bash
./ddueruem.py examples/xml/mendonca_dis.xml --svo pre_cl_size n:1
```
**WARNING** pre_cl_.. is only compatible with `.xml` files which are not sxfm ~~and force is only compatible with `.dimacs` files at the moment~~
Build BDD without svo, from order file
```bash
./ddueruem.py parsed/out/mendonca_dis-pre_cl_size-1.orders --bdd cudd lib_t:-1 dvo:off dump:True
```