/* We need to allocate one element extra, since we might use
* (unaligned) 4-wide SIMD loads to access rvec entries.
*/
- f.resize(top_global->natoms + 1);
+ f.resize(gmx::paddedRVecVectorSize(top_global->natoms));
/* Print to log file */
walltime_accounting_start(walltime_accounting);