Updated 03/26/2010

Installation:
===================================================
Download and unpack espresso-4.0.3 from the QE-forge 
website at http://www.pwscf.org.  Copy EPW/ into the 
espresso-4.0.3 directory, then execute the following 
commands from espresso-4.0.3/:
./configure
make ph
cd EPW/src
make 

This will create an epw.x executable in EPW/bin, which
can be used with pw.x and ph.x to run the examples in the
EPW/examples directory.

Known bugs/issues:
===================================================
When performing calculations on a system WIHTOUT 
inversion symmetry, the electron-phonon matrix 
elements for the -q of the star are not computed 
correctly.  Time reversal symmetry must not be
used.  Instead the coarse mesh phonons need to
be calculated explicitly for q and -q at certain
phonon wavevectors.

Ultrasoft pseudopotentials have not been tested, 
although they have been implemented.  Particularly
the wannier90 conversion of uspp wavefunctions 
seems to be incorrect.

Currently there is no parallelization over G-vectors.  
Practically this means that in order to run EPW 
in parallel,  NPOOL=NPROC must be satisfied. 


Troubleshooting:
===================================================
The program seems to hang on the "Running Wannier90" line
------------------------------------------------
As of version 1.1, W90 sends the STOP command when it 
encounters a fatal error, thus not exiting smoothly.  Check
the end of the prefix.wout for specific errors.

Creating kgmap seems to take forever
------------------------------------------------
Yes it does.  It is currenlty not parallelized. For a system
with over 10^5 G-vectors, the kgmap may take several hours.


The lambda's and linewidths of the phonon selfenergy
are identically zero
------------------------------------------------
Are you performing calculations on a metal?  Have you 
correctly set the 'nbndskip' parameter?  What is Ef
and the DOS at Ef?