[13] RFR(T): 8225475: Node budget asserts on x86_32/64

[Patric Hedlin]

Dear all,

I would like to ask for help to review the following _trivial_ 
change/update:

Issue: https://bugs.openjdk.java.net/browse/JDK-8225475

     Adjusting (ad-hoc) loop cloning size estimate to be more 
pessimistic in order
to more consistently overestimate the control/data merge effects from 
cloning.

Testing: Running the manifesting test-cases for a full week without 
failure (x64).
          Additional testing on x86 by Aleksey Shipilev. Thanks Aleksey.


Diff below.

Best regards,
Patric

-----8<-----

diff -r d4de822dc421 -r 1a22ea71d2c2 src/hotspot/share/opto/loopnode.cpp
--- a/src/hotspot/share/opto/loopnode.cpp    Fri Jun 28 21:41:04 2019 -0700
+++ b/src/hotspot/share/opto/loopnode.cpp    Tue Jun 25 11:43:36 2019 +0200
@@ -2494,9 +2494,11 @@
        }
      }
    }
-  // Add data (x1.5) and control (x1.0) count to estimate iff both are > 0.
+  // Add data and control count (x2.0) to estimate iff both are > 0. 
This is
+  // a rather pessimistic estimate for the most part, in particular for 
some
+  // complex loops, but still not enough to capture all loops.
    if (ctrl_edge_out_cnt > 0 && data_edge_out_cnt > 0) {
-    estimate += ctrl_edge_out_cnt + data_edge_out_cnt + 
data_edge_out_cnt / 2;
+    estimate += 2 * (ctrl_edge_out_cnt + data_edge_out_cnt);
    }

    return estimate;