Andrew Caldwell, Andrew B. Kahng Igor Markov |
I. | Introduction and overview |
II. | Partitioning Formats |
III. | Publicly available instances, solutions and reference performance results |
IV. | Executable Utilities (converters, generators, statistics browsers, evaluators, constraint verifiers) |
V. | Optimizers and other non-trivial executables |
VI. | Performance results |
VII. | Common in-memory representations, parsers and other source codes |
II. New Partitioning Formats
The proposed formats have been designed according to general guidelines for data formats in the GSRC bookshelf, and are intended for standardization within the partitioning and placement slot, replacing or coexisting with previously available formats. This slot will provide benchmark instances in the new formats, including instanes earlier available in other formats, as well as a variety of executables and source code for conversion, evaluation and optimization.
Netlists should be represented with .nodes, .nets and .wts files explained in the Hypergraph slot (follow links to "old fomats", i.e., no XML). In addition to those, .blk, .fix and .sol files can be used to represent balance and fixed node constraints for partitioning. Also see this small example.
V. Optimizers and other non-trivial executables
More information on MLPart is available here.
where filename.aux contains one line of the form
PartProb : ibm10.net ibm10.are 2way10%.blk ibm10.fix
or
PartProb : ibm10.nodes ibm10.nets ibm10.wts 2way10%.blk ibm10.fix
other supported options are
-help prints available options
-num ## each experiment contains ## independent starts with the best solution and total runtime as its result
-runs ## launches ## independent experiments and reports stats
-clip turns on CLIP FM (default is LIFO FM)
where file.aux contains one line of the form
PartProb : ibm10.net ibm10.are 2way10%.blk ibm10.fix
or
PartProb : ibm10.nodes ibm10.nets ibm10.wts 2way10%.blk ibm10.fix
other supported options are
-help prints available options
-num ## each experiment contains ## independent starts with the best solution and total runtime as its result
-runs ## launches ## independent experiments and reports stats
-skipVCycle turns on classic multi-level partitioning (faster)
VI. Performance results for IBM benchmarks: 2%, 10% (from the ASPDAC 2000 paper).
VII. Common in-memory representations, parsers and other source codes
UCLA/UMich PD tools release (see also the MLPart page)