[dpdk-dev] [RFC PATCH 0/1] eventtimer: introduce event timer wheel

[Jerin Jacob]

-----Original Message-----
> Date: Tue, 29 Aug 2017 17:48:58 +0200
> From: Thomas Monjalon <thomas at monjalon.net>
> To: Jerin Jacob <jerin.jacob at caviumnetworks.com>
> Cc: "Carrillo, Erik G" <erik.g.carrillo at intel.com>, dev at dpdk.org,
>  "Richardson, Bruce" <bruce.richardson at intel.com>, "Van Haaren, Harry"
>  <harry.van.haaren at intel.com>, "hemant.agrawal at nxp.com"
>  <hemant.agrawal at nxp.com>, "Eads, Gage" <gage.eads at intel.com>,
>  "nipun.gupta at nxp.com" <nipun.gupta at nxp.com>, "Vangati, Narender"
>  <narender.vangati at intel.com>, "Rao, Nikhil" <nikhil.rao at intel.com>,
>  "pbhagavatula at caviumnetworks.com" <pbhagavatula at caviumnetworks.com>,
>  "jianbo.liu at linaro.org" <jianbo.liu at linaro.org>, "rsanford at akamai.com"
>  <rsanford at akamai.com>
> Subject: Re: [dpdk-dev] [RFC PATCH 0/1] eventtimer: introduce event timer
>  wheel
> 
> 29/08/2017 17:41, Jerin Jacob:
> > From: Thomas Monjalon <thomas at monjalon.net>
> > > 25/08/2017 12:25, Jerin Jacob:
> > > > From: "Carrillo, Erik G" <erik.g.carrillo at intel.com>
> > > [...]
> > > > > In summary, it looks like our solutions align fairly well, and I propose that we take on the software implementation if there are no objections.
> > > > 
> > > > Sure, no objection.
> > > 
> > > Good to see such a basic function generalized for NPU and CPU.
> > > 
> > > Are you going to use rte_timer for CPU implementation?
> > > Does it mean that event_timer supersedes rte_timer?
> > 
> > IMO, we don't need to supersedes the rte_timer. The eventdev or event_timer is
> > an optional component. It is application decision to use poll mode vs event
> > driver model or combination of two.
> 
> Trying to make clear when using poll mode or event model,
> regarding CPU/NPU differences:
> If using poll mode + rte_timer, we cannot leverage NPU offloads.

Yes.

> If using event model, is it as much efficient for generic CPU?

It is case by case. I think, ATOMIC queues and PARALLEL queues can be
effectively implemented for the generic CPU. ORDERED queue(the reordering
business) may not very effective. So it is boiled down to application
requirements.