On Feb 12, 2014, at 4:12 PM, Pravin Shelar wrote:
> On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme
> wrote:
...
>> +#endif
>
> How much overhead do we have for non-NUMA systems without this compile
> time check.
> Since we always allocate node zero stat I do not think update stats
> has tha
On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme wrote:
> Keep kernel flow stats for each NUMA node rather than each (logical)
> CPU. This avoids using the per-CPU allocator and removes most of the
> kernel-side OVS locking overhead otherwise on the top of perf reports
> and allows OVS to scale b
Keep kernel flow stats for each NUMA node rather than each (logical)
CPU. This avoids using the per-CPU allocator and removes most of the
kernel-side OVS locking overhead otherwise on the top of perf reports
and allows OVS to scale better with higher number of threads.
With 9 handlers and 4 reval
