On Thu, Aug 17, 2017 at 4:04 PM, Luc Van Oostenryck <luc.vanoostenryck@xxxxxxxxx> wrote: > > Interesting and strange. > This doesn't really correspond to the numbers I have here. > Not at all, in fact, but I have only some rough measurements. > Just let me know if you need to get more number. My new server is itchy to compile stuff :-) The good thing about this server is that it does not do CPU frequency scale due to CPU over heating. So the number is very consistent. I can kind of see why your patch is slightly slow now. It allocate more phi nodes and remove them later. Also there is the overhead of the ptrmap. Both of those cause extra memory allocation, as we learn in the ptrlist node size case, it can slow things down a bit. Chris -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
