Skip to content

Improve join order for RangeAnalysis::boundedPhi #8393

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
smowton wants to merge 3 commits into from
Closed

Improve join order for RangeAnalysis::boundedPhi #8393

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
d9dfcb2
Improve join order for RangeAnalysis::boundedPhi
smowton Mar 10, 2022
c4aabce
Force CI
smowton Mar 10, 2022
8fbdfd5
Force CI
smowton Mar 10, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 15 additions & 1 deletion java/ql/lib/semmle/code/java/dataflow/RangeAnalysis.qll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -725,6 +725,16 @@ private predicate boundedPhiCandValidForEdge(
)
}

pragma[noinline]
private predicate boundedPhiCand2(
SsaPhiNode phi, Bound b, int delta, boolean upper, boolean fromBackEdge, int origdelta,
Reason reason
) {
exists(SsaVariable inp, SsaReadPositionPhiInputEdge edge | edge.phiInput(phi, inp) |
boundedPhiCandValidForEdge(phi, b, delta, upper, fromBackEdge, origdelta, reason, inp, edge)
)
}

/**
* Holds if `b + delta` is a valid bound for `phi`.
* - `upper = true` : `phi <= b + delta`
Expand All @@ -734,7 +744,11 @@ private predicate boundedPhi(
SsaPhiNode phi, Bound b, int delta, boolean upper, boolean fromBackEdge, int origdelta,
Reason reason
) {
forex(SsaVariable inp, SsaReadPositionPhiInputEdge edge | edge.phiInput(phi, inp) |
// Forcibly split into an existence and a forall stage instead of using forex because otherwise

@hvitved hvitved Mar 10, 2022

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am surprised to see that this doesn't use the "ranking approach" as in e.g. the modulus library. Replacing recursion through forall/forex with recursion over ranked indices typically performs much better.

@smowton smowton Mar 11, 2022

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Like #8405 ? Which do you prefer?

@hvitved hvitved Mar 14, 2022

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes. I prefer that PR over this one.

@hvitved hvitved Mar 14, 2022

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(this one also appears to have a significant slow-down on palatable/lambda).

// for certain seminaive variants we would try to postpone restricting by `edge.phiInput(phi, inp)`
// until too late in the process.
boundedPhiCand2(phi, b, delta, upper, fromBackEdge, origdelta, reason) and
forall(SsaVariable inp, SsaReadPositionPhiInputEdge edge | edge.phiInput(phi, inp) |
boundedPhiCandValidForEdge(phi, b, delta, upper, fromBackEdge, origdelta, reason, inp, edge)
)
}
Expand Down