Created attachment 649764 [details] [diff] [review]
This patch changes the recompile checks inserted by JM and Ion to use a slightly higher threshold for inner loops. The result is that we are more likely to enter outer loops instead of inner loops via OSR.
This closes about half the gap with JM+TI on ss-fannkuch (the rest seems to be compilation time, we are now faster than JM+TI on the modified benchmark in bug 777572), fixes the micro-benchmark in bug 779860 comment 0 and is a small v8-richards win.
*** Bug 779860 has been marked as a duplicate of this bug. ***