OS#15997192: Disable strict valuetype assert for hoistable syms #5547
Conversation
This file contains hidden or 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
The assert says that for a hoistable value, if its valuetype is primitive in the loop landing pad, it should be atleast LikelyPrimitive inside the loop.
This assert does not hold true in a few cases when we force specialize syms in the loop landing pad.
Ex :
landingPad:
a = v1;
b = v1;
loop :
b = v2; // forces float conversion in landing pad
c = a;
jump loop
During force specializing in loop landing pad of b, both a and b have the same valueinfo, so they transition to a DefiniteFloat.
This takes b to definite float in the loop landing pad but not inside the loop.
We do not have a way to force Specialize a in the loop, we cannot assume force specializing b again in the loop will update the valueInfo of a,
because they may have different value numbers in the loop. Since we don't have a way to lookup values by valueNumber, there is no easy way to fix this gap currently.
Added test to keep track of this gap.
rajatd
approved these changes
Jul 27, 2018
Collaborator
|
LGTM! |
LouisLaf
approved these changes
Jul 27, 2018
Contributor
Author
|
@dotnet-bot test Windows 10 ci_slow_x64_debug |
chakrabot
pushed a commit
that referenced
this pull request
Jul 27, 2018
… for hoistable syms Merge pull request #5547 from meg-gupta:disableassert The assert says that for a hoistable value, if its valuetype is primitive in the loop landing pad, it should be atleast LikelyPrimitive inside the loop. This assert does not hold true in a few cases when we force specialize syms in the loop landing pad. Ex : landingPad: a = v1; b = v1; loop : b = v2; // forces float conversion in landing pad c = a; jump loop During force specializing in loop landing pad of b, both a and b have the same valueinfo, so they transition to a DefiniteFloat. This takes b to definite float in the loop landing pad but not inside the loop. We do not have a way to force Specialize a in the loop, we cannot assume force specializing b again in the loop will update the valueInfo of a, because they may have different value numbers in the loop. Since we don't have a way to lookup values by valueNumber, there is no easy way to fix this gap currently. Added test to keep track of this gap.
chakrabot
pushed a commit
that referenced
this pull request
Jul 27, 2018
…luetype assert for hoistable syms Merge pull request #5547 from meg-gupta:disableassert The assert says that for a hoistable value, if its valuetype is primitive in the loop landing pad, it should be atleast LikelyPrimitive inside the loop. This assert does not hold true in a few cases when we force specialize syms in the loop landing pad. Ex : landingPad: a = v1; b = v1; loop : b = v2; // forces float conversion in landing pad c = a; jump loop During force specializing in loop landing pad of b, both a and b have the same valueinfo, so they transition to a DefiniteFloat. This takes b to definite float in the loop landing pad but not inside the loop. We do not have a way to force Specialize a in the loop, we cannot assume force specializing b again in the loop will update the valueInfo of a, because they may have different value numbers in the loop. Since we don't have a way to lookup values by valueNumber, there is no easy way to fix this gap currently. Added test to keep track of this gap.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
The assert says that for a hoistable value, if its valuetype is primitive in the loop landing pad, it should be atleast LikelyPrimitive inside the loop.
This assert does not hold true in a few cases when we force specialize syms in the loop landing pad.
Ex :
landingPad:
a = v1;
b = v1;
loop :
b = v2; // forces float conversion in landing pad
c = a;
jump loop
During force specializing in loop landing pad of b, both a and b have the same valueinfo, so they transition to a DefiniteFloat.
This takes b to definite float in the loop landing pad but not inside the loop.
We do not have a way to force Specialize a in the loop, we cannot assume force specializing b again in the loop will update the valueInfo of a,
because they may have different value numbers in the loop. Since we don't have a way to lookup values by valueNumber, there is no easy way to fix this gap currently.
Added test to keep track of this gap.