Address recent performance regression #90
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Reported in the schrodigner bug tracker as SHARED-7789 Our performance monitoring tests noticed that coordgen got about 10x slower after PR schrodinger#81. I guess the inner loop of minimization must be hitting the max pretty frequently! We should address that too. (this doubles the max iterations from before schrodinger#81, but that's ok because schrodinger#81 halved the calls to this function) Before this change: slowest: 1.442s average: 0.044s molecules above 0.1s: 566 molecules above 1s: 9 after this change: slowest: 0.151s average: 0.006s molecules above 0.1s: 12 molecules above 1s: 0
cdvonbargen
approved these changes
Apr 9, 2021
rachelnwalker
approved these changes
Apr 9, 2021
| @@ -37,7 +37,7 @@ static const unsigned int MAXIMUM_NUMBER_OF_SCORED_SOLUTIONS = 10000; | |||
| static const float REJECTED_SOLUTION_SCORE = 99999999.f; | |||
| CoordgenMinimizer::CoordgenMinimizer() | |||
| { | |||
| m_maxIterations = 10000; | |||
There was a problem hiding this comment.
Was this just a typo? Or were the previous changes expected to make the minimizer converge much sooner than 10000 iterations?
There was a problem hiding this comment.
It was both! Nic didn't expect that so many structures hit the max, so thought that this change wouldn't matter.
And I, in reviewing it, read it as 1,000 instead of 10,000. (a "read-o", I guess)
d-b-w
added a commit
to d-b-w/rdkit
that referenced
this pull request
Apr 9, 2021
This is pretty urgent, because it includes a fix for a 10X performance regression in coordgen (schrodinger/coordgenlibs#90). It also fixes coordinate generation for 2.2.2 bicyclic systems.
greglandrum
pushed a commit
to rdkit/rdkit
that referenced
this pull request
Apr 10, 2021
This is pretty urgent, because it includes a fix for a 10X performance regression in coordgen (schrodinger/coordgenlibs#90). It also fixes coordinate generation for 2.2.2 bicyclic systems.
greglandrum
pushed a commit
to rdkit/rdkit
that referenced
this pull request
May 12, 2021
This is pretty urgent, because it includes a fix for a 10X performance regression in coordgen (schrodinger/coordgenlibs#90). It also fixes coordinate generation for 2.2.2 bicyclic systems.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reported in the schrödinger bug tracker as SHARED-7789
Our performance monitoring tests noticed that coordgen got about
10x slower after PR #81. I guess the inner loop of minimization
must be hitting the max pretty frequently! We should address
that too.
(this doubles the max iterations from before #81, but that's ok
because #81 included a tidy refactor that halved the calls to this
function)
Before this commit:
After this commit: