-
Enhancement
-
Resolution: Unresolved
-
P4
-
21
This includes the following steps:
See if we can do more verification for LoadNode::Value, maybe disable some optimizations during verification?
JDK-8302071 IGVN shuffle_worklist: we should try to do random selection instead!
Take more steps to do similar checks for Node::Ideal and Node::Identity.
JDK-8308149 Verify number of out-projection from MultiBranch
Verify consistency of Loop nodes (especially CountedLoop with tripcount phi etc)
JDK-8293362: Extend detection of modified nodes to incremental inlining
Investigate what to do about JDK-8273496
After all the verification is added, we should do this:
Refactor the CCP / IGVN notification code JDK-8298094
- relates to
-
JDK-8273496 Make sure that CCP adds nodes to the IGVN worklist
- Open
-
JDK-8293362 Extend detection of modified nodes to incremental inlining
- Open
-
JDK-8298094 Refactor PhaseCCP and IGVN notification
- Open
-
JDK-8308149 C2 VerifyIterativeGVN: verify CFG consistency (MultiBranchNode::required_outcnt() == outcnt())
- Open
-
JDK-8302071 StressIGVN: not just initial shuffle_worklist: always pick next node at random
- In Progress
-
JDK-8298952 All nodes should have type(n) == Value(n) after IGVN
- Resolved
-
JDK-8343067 C2: revisit constant-offset AddP chains after successful input idealizations
- Resolved
-
JDK-8257197 Add additional verification code to PhaseCCP
- Resolved