OpenMPL stands for open multiple patterning lithography framework.

Stitch Insertion	Graph Simplification	Decomposition

• GCC

  • Recommend GCC 4.8 or later.
  • Other compilers may also work, but not tested.

• CMake

  • Require 3.8.2 or later.

• Boost

  • Require 1.55...1.72.
  • Need to install and visible for linking.
  • Custom installation path may require to export BOOST_ROOT for CMake.

• Limbo

  • Integrated as a git submodule.

• Gurobi (Optional)

  • ILP solver.

• Wei Li, Yuzhe Ma, Qi Sun, Lu Zhang, Yibo Lin, Iris Hui-Ru Jiang, Bei Yu, David Z. Pan, “OpenMPL: An Open Source Layout Decomposer”, accepted by IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems (TCAD). (preprint)

• Wei Li, Yuzhe Ma, Qi Sun, Yibo Lin, Iris Hui-Ru Jiang, Bei Yu, David Z. Pan, "OpenMPL: An Open Source Layout Decomposer", IEEE International Conference on ASIC (ASICON), Chongqing, China, Oct. 29–Nov. 1, 2019. (preprint) (slides)

We provide the docker version to avoid annoying compilation. You can pull the image by

$ docker pull wadmes/openmpl:latest

The OpenMPL folder locates in /home/OpenMPL

$ git clone https://github.com/limbo018/OpenMPL.git 
$ cd OpenMPL
$ git submodule update --init --recursive
$ mkdir build
$ cd build
$ cmake .. 
$ make
$ make install

The default installation path is the repo directory. It can be changed via

cmake .. -DCMAKE_INSTALL_PREFIX=your_installation_path

• Contact or metal layer decomposition
• Stitching
• Support 3 or 4 coloring
• Density control
• Multi-threading
• Small memory usage
• Multiple algorithms:
  • ILP (Gurobi API)
  • SDP (Csdp API)
  • LP (Gurobi API)
  • Dancing Links

The Csdp API used in OpenMPL has been modified and built for threading safety at high level.

Small benchmarks (ISCAS) Download big benchmarks (ISPD19) Download

$ cd bin/
$ ./OpenMPL

A table of options :

-help (false)                toggle printing help message
-in                          input gds file name
-out ()                      output gds file name
-coloring_distance (0)       a floating point number indicating number of coloring distance in nanometer
-color_num                   an integer indicating number of masks (colors) < 3|4 >
-simplify_level (3)          an integer indicating graph simplification level < 0|1|2|3 >
-thread_num (1)              an integer indicating maximum thread number
-path_layer                  an integer indicating layer for conflict edges
-precolor_layer              an integer indicating layer for pre-colored patterns
-uncolor_layer               an integer indicating layer for coloring
-algo (BACKTRACK)            algorithm type < ILP|BACKTRACK|LP|SDP >
-shape (RECTANGLE)           shape mode < RECTANGLE|POLYGON >
-verbose (false)             toggle controlling screen messages
-dbg_comp_id (4294967295)    debug component id
-use_stitch                  use stitch to avoid conflict
-gen_stitch                  generate stitch candidate
-weight_stitch               a floating point number indicating the weight of stitch                     

One exmaple : /bin/run.sh. Another example for large benchmark: /bin/run_big.sh

• default CFLAGS of boost and gurobi are different in newest version

• downgrade the boost version

• SimpleMPL.cpp:461:5: error: ‘graph_simplification_type’ has no member named ‘set_isVDDGND’
  

  • checkout to stitch branch in your limbo directory

• BSD-3-clause License [LINK]