You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If I open the sequences-18.log and then apply the filter "Show quant k!354" then the tool gets stuck in the VisibleInstGraph::reconnect function, see here.
I've tried to limit the maximum number of edges to 1000 or so to inspect the intermediate graph of quant k!354 and saw that it adds a lot of edges, see screenshot below.
I'll also link two SVG files to inspect the subgraph of k!354 (dark purple nodes), once without and once with other quants:
My suspicion is that it's not a correctness issue but a performance issue, due to how the equality nodes are represented in the edges of the reconnected graph.
The text was updated successfully, but these errors were encountered:
oskari1
changed the title
Slow reconnect in sequences-18.log
Reconnect gest stuck in sequences-18.log
Apr 22, 2024
oskari1
changed the title
Reconnect gest stuck in sequences-18.log
Reconnect stuck in sequences-18.log
Apr 22, 2024
where if only the V nodes are visible, then one gets 2^3 indirect edges (grows exponentially with the hidden length). I think that we should change the reconnect to only consider immediate hidden children/parents, without enumerating all the paths between (such that there would only be 2 indirect edges in the above graph, for the 2 hidden children * 1 hidden parent)
If I open the sequences-18.log and then apply the filter "Show quant k!354" then the tool gets stuck in the
VisibleInstGraph::reconnect
function, see here.I've tried to limit the maximum number of edges to 1000 or so to inspect the intermediate graph of quant k!354 and saw that it adds a lot of edges, see screenshot below.
I'll also link two SVG files to inspect the subgraph of k!354 (dark purple nodes), once without and once with other quants:
My suspicion is that it's not a correctness issue but a performance issue, due to how the equality nodes are represented in the edges of the reconnected graph.
The text was updated successfully, but these errors were encountered: