fix(ivy): avoid type coercion in binding-related instruction asserts #29470
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
|
Profile before: Profile after: |
pkozlowski-opensource
force-pushed
the
ivy_perf_material_demo
branch
from
43e3ee7
to
72f728e
Compare
pkozlowski-opensource
added
action: review
kara
requested changes
Mar 22, 2019
packages/core/src/util/assert.ts
Outdated
| @@ -23,7 +23,7 @@ export function assertEqual<T>(actual: T, expected: T, msg: string) { | |||
| } | |||
|
|
|||
| export function assertNotEqual<T>(actual: T, expected: T, msg: string) { | |||
| if (actual == expected) { | |||
| if (actual === expected) { | |||
There was a problem hiding this comment.
We have an assertEqual (==) and an assertSame (===), so to be consistent, the inverse of === should be assertNotSame. At the least, we should update the name, but we probably need a new assert function for this?
Ha ha, @kara you are totally right :-) I will fix it - thnx for spotting! |
kara
added
action: cleanup
ivy's bindingUpdated instruction is using the assertNotEqual check to make sure that NO_CHANGE value (of type Object) is not passed as a value to be dirty-checked. In practice it means that any value passed as a binding value would be compared to the NO_CHANGE object. It turns out that the assertNotEqual is using == and given that binding values are of different type and we always compare it to the NO_CHANGE object we were doing lots of type coercion. It resulted in calls to expensive types conversions and calls to Object.toString(). A profiler reported ~15% of the self time spent in the assertNotEqual but it turns out that removing type coercion speeds up Material Chips with input scenario much more (~40ms down to ~20ms). This PR introduces new assert method `assertNotSame` that uses strict equality check. The new assertion is used in binding instructions to compare to NO_CHANGE object reference.
pkozlowski-opensource
force-pushed
the
ivy_perf_material_demo
branch
from
72f728e
to
fa37ff9
Compare
pkozlowski-opensource
removed
the
action: cleanup
|
@kara added a new |
kara
added
action: merge
wKoza
pushed a commit
to wKoza/angular
that referenced
this pull request
Apr 17, 2019
…ngular#29470) ivy's bindingUpdated instruction is using the assertNotEqual check to make sure that NO_CHANGE value (of type Object) is not passed as a value to be dirty-checked. In practice it means that any value passed as a binding value would be compared to the NO_CHANGE object. It turns out that the assertNotEqual is using == and given that binding values are of different type and we always compare it to the NO_CHANGE object we were doing lots of type coercion. It resulted in calls to expensive types conversions and calls to Object.toString(). A profiler reported ~15% of the self time spent in the assertNotEqual but it turns out that removing type coercion speeds up Material Chips with input scenario much more (~40ms down to ~20ms). This PR introduces new assert method `assertNotSame` that uses strict equality check. The new assertion is used in binding instructions to compare to NO_CHANGE object reference. PR Close angular#29470
|
This issue has been automatically locked due to inactivity. Read more about our automatic conversation locking policy. This action has been performed automatically by a bot. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
ivy's bindingUpdated instruction is using the assertNotEqual check to make
sure that NO_CHANGE value (of type Object) is not passed as a value to be
dirty-checked. In practice it means that any value passed as a binding
value would be compared to the NO_CHANGE object.
It turns out that the assertNotEqual was using == instead of === and given
that binding values are of different type and we always compare it to the
NO_CHANGE object we were doing lots of type coercion. It resulted in calls
to expensive types conversions and calls to Object.toString().
A profiler reported ~15% of the self time spent in the assertNotEqual
but it turns out that removing type coercion speeds up Material Chips with
input scenario much more (~40ms down to ~20ms).