On Thu, 24 Oct 2013 19:23:55 -0400 John Scipione <jscipione@xxxxxxxxx> wrote: > On Thu, Oct 24, 2013 at 7:11 PM, Pawel Dziepak <pdziepak@xxxxxxxxxxx> wrote: > > Since the scheduler code has grown quite big I am trying to remove as > > much unnecessary logic as possible. The thing that caught my attention > > is the possibility to disable a logical processor. > > > > As far as I understand there are two undocumented (i.e. BeBook doesn't > > mention them but some old apps, e.g. Pulse, use them) syscalls > > set_cpu_enabled and cpu_enabled that allow disabling/enabling a CPU > > and check whether CPU is enabled. I don't really see why anyone would > > like to use that. Definitely not for performance reasons. If because > > of the energy consumption then the new scheduler has power saving mode > > with small task packing which tries to keep as little CPUs running as > > possible. > > > > So, unless there is any situation it may be useful to manually disable > > CPU that I am missing, I am going to remove that feature (the menu in > > Deskbar for disabling CPUs could be replaced by menu for choosing the > > scheduler operating mode once public API for that is done). > > _kern_set_cpu_enable and _kern_cpu_enable have to stay tough (unless > > my assumption that they are part of the original BeOS ABI is wrong) > > but can be implemented totally in libroot returning respectively B_OK > > and true without even calling the kernel. > > > > Anyone going to miss that feature? > > It's a neat feature allowing you to test the system under load by > dynamically turning a pocessors on and off to see the difference. I > would definitely miss the feature if it were removed. +1. I would definitely miss it. While it's usefulness is limited, it is almost always the first thing I demo to techie people (and can wow them with) So... -1 on removing it. -- Alex