On Sat, Feb 6, 2016, 6:06 AM <[email protected]> wrote:
> [email protected] > <https://groups.google.com/forum/?utm_source=digest&utm_medium=email#!forum/mechanical-sympathy/topics> > Google > Groups > <https://groups.google.com/forum/?utm_source=digest&utm_medium=email/#!overview> > <https://groups.google.com/forum/?utm_source=digest&utm_medium=email/#!overview> > Topic digest > View all topics > <https://groups.google.com/forum/?utm_source=digest&utm_medium=email#!forum/mechanical-sympathy/topics> > > - Static branch prediction (was: On intel x86 processors do you have > branch predictions on fixed loops ?) > <#m_4640073617864472708_group_thread_0> - 1 Update > > Static branch prediction (was: On intel x86 processors do you have branch > predictions on fixed loops ?) > <http://groups.google.com/group/mechanical-sympathy/t/a4c37a4db5f03947?utm_source=digest&utm_medium=email> > Matt Godbolt <[email protected]>: Feb 05 02:58PM > > Hi Matt, > > A quick reply: thanks again for the link! My reading list is significantly > backed up now :) > > For now I'm going to concentrate on honing my technique and scripts on the > Arrendale - mainly as it's a computer I have available even when I'm > commuting (which is where I get most of the time for this!). Once I'm > comfortable explaining branch prediction there, I'll move on to the Sandy, > Ivy and Haswell: each of which has significant BP improvements. > > I'm also first starting on interpreting the BTB size - under the assumption > that there's no taken/not-taken information stored there; and that a miss > in the BTB always causes one of the "BP failure" counters to tick up. At > least on Arrendale/Westmere, there's a bunch of counters for this; early > ones where the decoder spots a branch the BTB failed to notice, for > example. I have a ton of notes and thoughts; I'll probably write up > something over the next few days and post it here (or a link to a blog post > or similar). > > Thanks, Matt (G) :) > > Back to top <#m_4640073617864472708_digest_top> > You have received this digest because you're subscribed to updates for > this group. You can change your settings on the group membership page > <https://groups.google.com/forum/?utm_source=digest&utm_medium=email#!forum/mechanical-sympathy/join> > . > To unsubscribe from this group and stop receiving emails from it send an > email to [email protected]. > -- You received this message because you are subscribed to the Google Groups "mechanical-sympathy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web, visit https://groups.google.com/d/msgid/mechanical-sympathy/CAJxDUeW89CD4K6o4BVMSaYWRoTrn6J9QXEVL%2BWaPm1iEK2eVLw%40mail.gmail.com.
