<mattip>
I wonder if that might be connected to the numpy and winconsoleio problems
jacob22 has joined #pypy
smarr has joined #pypy
<smarr>
just to follow on from yesterday. I moved my "function entry" jitdriver a little later in the code to have just a single `jit_merge_point` again, but it basically doesn't change things, still major slow downs on recursive benchmarks.
<smarr>
Is there anything general I should be looking out for when it comes to getting good traces on recursive programs? (in terms of jitdriver placement/configuration)