-
Notifications
You must be signed in to change notification settings - Fork 4.9k
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.
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
Stop producing and handling ADDR
nodes
#78246
Conversation
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch Issue DetailsMostly just dead code and cases where we can use Some positive diffs are expected from the MD array import change.
|
8d332c6
to
bcef6eb
Compare
d0d8618
to
bcef6eb
Compare
f792610
to
99c19dd
Compare
ADDR
-related codeADDR
nodes
ADDR
nodesADDR
nodes
d09f03c
to
5b22300
Compare
ca22141
to
e62d5fc
Compare
e62d5fc
to
71b97d0
Compare
@dotnet/jit-contrib We are getting eerily close to |
/azp run runtime-coreclr jitstress, runtime-coreclr libraries-jitstress |
Azure Pipelines successfully started running 2 pipeline(s). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Changes LGTM. Thanks for your continuing valuable contributions!
I kicked off stress runs.
libraries-jitstress had a few failures: one known, one looked like infra, Linux/arm tailcallstress had a "new" crash in System.Memory.Tests. Set to rerun to see if it repros. |
Didn't repro; all failures are "known". |
With this change, the
GT_ADDR
oper (for a short time before its eventual deletion) becomes "synthetic": no actual nodes of it will be created.We also start using local address nodes in more places and delete dead code.
Diffs: minor GC info size reductions (
LCL_VAR_ADDR
s are typedI_IMPL
); some substitution diffs (due to fewerGLOB_REF
s).There are nice TP improvements:
0.3%
on 64 bit platforms,1%
on 32 bit platforms, of which the latter comes mostly from theOperIsIndir
fix.