Francesco Pietra <[EMAIL PROTECTED]>: > int2e_buf_read:asynch IO failed 1301
The only references i could find in the net are like <http://www.emsl.pnl.gov/docs/nwchem/nwchem-support/2006/03/0047._NWCHEM_int2e_buf_read> > ---Cleaning the swap from time to time? Normally kernel and initscripts care for these things. But assumed a HPC system would not boot for many months. ...then what? Swap swap partitions on the fly ? (with one spare) But anyway, swap is not the asnwer. Depending on your system and applications you may consider a special high performace linux kernel, or at least compiler your own version of vanilla, there are some useful options. > ---Swap size with respect to ram size? I can't think of any important reason, besides hibernation-to-swap, which most probably would not apply to your system. Figure what you want or need. Swap is real slow, but could be used as kind of emergency trigger (providing the time for reaction) m° (and maybe http://www.llnl.gov/linux/ucrl-jc-153559.html#1 http://www.nec.com.au/products_detail.aspx?view=146 http://www.csm.ornl.gov/pvm/ )
