Sounds like getting the hardware information from the vendors should help. There are some indications that we missed the train, but if there is another train about to be boarded, it'd be good to be on it if it makes sense :-)
Milan On 2013-02-25, at 3:18 PM, Jet Villegas <j...@mozilla.com> wrote: > I see what you mean there, and I don't mean to trivialize the driver details. > I think you can get past much of that by reducing the number of devices to > support (eg. constrain to one B2G partner device initially,) and reducing the > cost to stand up the backend. I don't know how far Cairo got to getting it > integrated, but that may be sufficient to test if it has potential for Gecko. > Vlad wrote it, so we should ask him. If that looks good, we can see about > wiring it up in Azure. I'm all for using the cheapest, hackiest way to find > out if these *experiments* can bear fruit, before deciding on committing to > full-blown implementations. > > --Jet > > > ----- Original Message ----- > From: "Nicholas Cameron" <nick.r.came...@gmail.com> > To: "mozilla dev platform" <mozilla.dev.platf...@googlegroups.com> > Cc: dev-platform@lists.mozilla.org, "Jet Villegas" <j...@mozilla.com> > Sent: Monday, February 25, 2013 11:40:40 AM > Subject: Re: OpenVG Azure backend > > On Tuesday, February 26, 2013 8:34:39 AM UTC+13, Jet Villegas wrote: >> AIUI, a key goal of our Layers Refactoring is to significantly reduce the >> cost of standing up new GFx backends (the first new one being D3Dv11.) As >> that progresses, I'd like to see detailed documentation about how new >> backends are added. I think we can get an intern or volunteer to stand up >> experimental backends and come with the perf numbers to inform our >> decisions. We've long predicted an increase in the number of low-level GFx >> driver surfaces we'll need to support, given our Android and FirefoxOS >> commitments. Let's make use of the infrastructure we're building to ensure >> that we can scale to support that. >> > > This would be an Azure backend, not a layers backend, so is not affected by > the refactoring. Implementing Azure backends is pretty nice (that was the > point of Azure), but that kind of thing is intrinsically quite a large amount > of work. Also, in general, implementing the backend is trivial compared to > dealing with the driver bugs that are exposed by it and general support costs. > _______________________________________________ > dev-platform mailing list > dev-platform@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-platform _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
