> Andi Kleen (iirc) says that non-temporal stores seem to be > big win in microbenchmarks (and I second that), but they are > a net loss when we are going to use zeroed page just after > zeroing. He recommends avoid using non-temporal stores
The rule of thumb is to only use non temporal stores when your data set is bigger than the L2/L3 caches of the CPU. This means >1MB. The kernel normally never works on data sets that big. For Christophers new background cleaner daemon it may be worth it when the queue is a LILO. This means it is likely there is a relatively long time between the clearing operation and a workload using it. But even then it is a very close call and would need clear benchmark numbers in macrobenchmarks. -Andi - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
