-
Notifications
You must be signed in to change notification settings - Fork 4.7k
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
Delete GT_INDEX
#69917
Delete GT_INDEX
#69917
Conversation
Tagging subscribers to this area: @JulieLeeMSFT Issue DetailsInstead: This deletes two usages of It also deletes a case of us creating layout-less
|
35dfb9f
to
2bdb2c0
Compare
@dotnet/jit-contrib |
65a9e57
to
ca9c61a
Compare
/azp run runtime-coreclr jitstress |
Azure Pipelines successfully started running 1 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.
This looks great. I kicked off jitstress to verify with a larger test set.
It is effectively no-CSE already because of how "optIsCSECandidate" works.
Instead: 1) For "ldelem", import "IND/OBJ(INDEX_ADDR)". 2) For "ldelema", import "INDEX_ADDR". This deletes two usages of "ADDR": 1) "OBJ(ADDR(INDEX))" from "ldelem<struct>". 2) "ADDR(INDEX)" from "ldelema".
Remove references to things that no longer exist.
Force-pushed to resolve a conflict with #70265. |
ca9c61a
to
c7bc540
Compare
Instead:
1) For
ldelem
, importIND/OBJ(INDEX_ADDR)
.2) For
ldelema
, importINDEX_ADDR
.This deletes two usages of
ADDR
:1)
OBJ(ADDR(INDEX))
fromldelem<struct>
.2)
ADDR(INDEX)
fromldelema
.It also deletes a case of us creating layout-less
IND struct
nodes and some amount of auxiliary code.Diffs: a few in the tests reflecting forward substitution's limit of the amount of nodes it is willing to forward, plus one in libraries tests stemming from this code only looking at
ADDR
s (fix in progress: #69961). There is also a fairly small TP regression (curiously, 64 bit-only) that has been payed for with the improvements from #69895.