decreasing #define HZ in the linux kernel for CPU/memory bound apps ?
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Josip Loncaric josip at icase.eduTue Apr 16 12:58:40 PDT 2002
- Previous message: decreasing #define HZ in the linux kernel for CPU/memory bound apps ?
- Next message: decreasing #define HZ in the linux kernel for CPU/memory bound apps ?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Greg Lindahl wrote: > > As for "what HZ would be a good value?", Alpha has always used 1000, > and it isn't a significant performance hit. But x86 started life on > much slower machines, and now we're stuck with 100, unless you want to > rebuild ALL your packages. > > I suspect IA64 uses 100 for compatibility reasons. I wonder how the > x86 emulator on AlphaLinux got around this... hm... A minor correction: HZ=1024 on Alphas and on ia64 (elsewhere HZ=100). HZ=1024 helps, e.g. it prevents certain kinds of timer-resolved TCP stalls in kernel 2.2 on Alphas. However, recompiling user programs which were built with HZ=100 would be a pain... and one might uncover new problems with the i386 hardware which has not been tested much with HZ=1024. Sincerely, Josip -- Dr. Josip Loncaric, Research Fellow mailto:josip at icase.edu ICASE, Mail Stop 132C PGP key at http://www.icase.edu./~josip/ NASA Langley Research Center mailto:j.loncaric at larc.nasa.gov Hampton, VA 23681-2199, USA Tel. +1 757 864-2192 Fax +1 757 864-6134
- Previous message: decreasing #define HZ in the linux kernel for CPU/memory bound apps ?
- Next message: decreasing #define HZ in the linux kernel for CPU/memory bound apps ?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
