Test new DCSH variable orderer to ensure it matches the paper performance

I think more testing is needed. I don't yet know whether the improvements we get in terms of better variable orderings and reduced synthesis effort match what the DCSH paper indicates. That would be good to check before lifting the 'experimental' label for end users to use it.

added CIF TypeEnhancement labels

marked this issue as related to #375 (closed)

This can benefit from the scripts of #375 (closed).

mentioned in issue #377 (closed)

marked this issue as related to #377 (closed)

mentioned in issue #378 (closed)

marked this issue as related to #378 (closed)

mentioned in issue #379 (closed)

marked this issue as related to #379 (closed)

mentioned in issue #380 (closed)

marked this issue as related to #380 (closed)

mentioned in merge request !327 (closed)

assigned to @ddennis

changed milestone to %v0.6

Initial results are not encouraging.

festo benchmark:

nodes, ops, setting
35467, 3003772, default
146337, 52115562, default+DCSH

A >4x increase in max nodes and >17x increase in number of operations, for enabling DCSH on top of the default settings.

waterway_lock benchmark:

nodes, ops, setting
130087, 76106672, default
508545, 1022514343, default+DCSH

A ~4x increase in max nodes and >13x increase in number of operations, for enabling DCSH on top of the default settings.

I'll have to investigate further.

mentioned in issue #196 (closed)

changed milestone to %v0.7

unassigned @ddennis

removed milestone %v0.7

I'll have to investigate further.

Clearly it was caused by not using linearized hyper-edges at that time.

@sthuijsman will rerun the experiments from the earlier paper for a new paper. We'll then see how it turns out.

Edited: @-reference had a typo

We are no longer reproducing the earlier paper in our new paper, but doing new experiments with current CIF. Therefore, it is no longer the goal to be like the old paper. We also can't be the same anymore, as we've made various other changes even as well. I'm closing this issue.

changed milestone to %v0.9

assigned to @ddennis

changed milestone to %v0.10

changed milestone to %v0.9

closed