Hello,

> With the io-c branch I see much better unpickling performance than
> before. But it still seems to be around 2 or 3 times slower than with
> cPickle in 2.6.

It's much closer here.

With 2.7 (trunk) and cPickle:
0.439934968948
0.669679880142
With the io-c branch:
0.136367082596
0.798221111298

> Is this expected at this point of io-c development?

20-30% slower is expected for binary I/O (which is what I get above),
not 2x or 3x slower.
What is your system?