Hi Jin,we had the same problem a long time ago (I think it was still HTS 1.1, never tried with 2.x),
in our case, p3 / 512M was not enough and after a few hours, the process started to intensively use the swap in a very (very) slow way (after a complete week-end, it was still in the middle of the training and we stopped it)
on differents 1024M systems (p4, centrino) we had no problem (even on a win32+cygwin it worked)
Alexis Jin K. Park wrote :
I am new to this HTS group.Can you give me the specification of the system that ran HTS-demo_CMU-ARCTIC-SLT. On the INSTALL file, it says 6-12 hours but on mine, Intel celeron/512M/Linux 2.6.9-55.EL/i686, it is running more than two days now.Maybe I should run on different machine.