On 10 March 2017 at 22:51, Jeff Hammond <[email protected]> wrote:
> > > On Fri, Mar 10, 2017 at 9:09 AM, Barry Smith <[email protected]> wrote: > >> >> > On Mar 10, 2017, at 9:58 AM, Lisandro Dalcin <[email protected]> wrote: >> > >> > On 10 March 2017 at 03:19, Barry Smith <[email protected]> wrote: >> > PetscMallocRegister(malloc,realloc,free,&basicmalloc); >> > PetscMallocRegister(PetscMallocDebug,PetscReallocDebug,Pets >> cFreeDebug,&debugmalloc); >> > PetscMallocRegister(PetscMallocHBW,PetscReallocHBW, >> PetscFreeHBW,&hbwmalloc); >> > >> > I'm wondering whether we should also handle calloc() here, and >> implement PetscCalloc properly. As I understand it, calloc != >> malloc+memset, and the differences might be important in >> multicore+multithreading scenarios and the first-touch policy. >> >> > My intuition is that any HPC code that benefits from mapping the zero page > vs memset is doing something wrong. > > I have the same feeling, but people will continue to use OpenMP no matter how much you advocate against it. -- Lisandro Dalcin ============ Research Scientist Computer, Electrical and Mathematical Sciences & Engineering (CEMSE) Extreme Computing Research Center (ECRC) King Abdullah University of Science and Technology (KAUST) http://ecrc.kaust.edu.sa/ 4700 King Abdullah University of Science and Technology al-Khawarizmi Bldg (Bldg 1), Office # 0109 Thuwal 23955-6900, Kingdom of Saudi Arabia http://www.kaust.edu.sa Office Phone: +966 12 808-0459
