Messages in this thread | | | Date | Tue, 12 Feb 2013 18:33:42 +0100 | Subject | Re: [PATCH 0/2] perf stat: add per-core count aggregation | From | Stephane Eranian <> |
| |
On Tue, Feb 12, 2013 at 6:29 PM, Andi Kleen <ak@linux.intel.com> wrote: >> >> > The idea itself is useful. >> > >> Yes, it is. > > BTW it would be even more useful if it could print some of the > statistics turbostat does (in particular frequency and C0 residency) > Often you only care about cycles not idle, and the frequency > tells you how fast the cycles happen. > You get C-state residency at the socket level with my perf uncore patch.
This patch is just about helping with unbalance at the physical core level.
> I think Cx could be added relatively easily as a software event, > but frequency doesn't fit very well into the perf counting model, > as it's really sampling. > > -Andi
| |