intel Iris 6200

Post requests to add new GPUs to the official whitelist here.

Moderators: Site Moderators, PandeGroup

Re: intel Iris 6200

Postby foldy » Tue Jan 03, 2017 5:51 pm

Don't get mad with bruce's pragmatic answers. It is the scientists from stanford pande group which decide what to do next. They use their time e.g. to extend some protein force fields implementation because that is super important for them. And this is why we have no intel iGPU fahcore or Mac OSX GPU core today.

BUT because of the more generic OpenCL implementation of the FahCores it may be possible to provide FahCores for intel iGPUs or Mac OSX without too much cost and gain 1%-10% more pflops. So if not any protein force field things block the pande group developers they might start testing these FahCores and we get some beta work units and maybe by end of the year the FahCores for intel iGPUs or Mac OSX are released to public.

The history of folding@home or other projects on BOINC shows that sometimes this plan does not come true.

As donors we wish that all our provided hardware can be used to the max (or 98% so that we can still surf the web and watch videos concurrently)
foldy
 
Posts: 853
Joined: Sat Dec 01, 2012 3:43 pm

Re: intel Iris 6200

Postby 7im » Tue Jan 03, 2017 6:06 pm

Research is their first priority, development falls below that, so they tend to focus on the lowest hanging and largest fruit on the performance tree. They have to ignore the fruit that has been sitting on the ground for too long as well. 15+ years of watching how the project functions gives bruce very accurate (though somewhat anecdotal) insight in to how a request like this is handled.

A request to support hardware "X" is what is consider a perpetual request. It's made often, and repeatedly, and given the about the same response, depending on feasibility. Let me see if I can get this about right... "Pande Group has been watching the development of hardware "X" closely. The cost/benefit analysis doesn't come out in our favor yet, but we will review this hardware again at a future date."
How to provide enough information to get helpful support
Tell me and I forget. Teach me and I remember. Involve me and I learn.
User avatar
7im
 
Posts: 15237
Joined: Thu Nov 29, 2007 4:30 pm
Location: Arizona

Re: intel Iris 6200

Postby ComputerGenie » Tue Jan 03, 2017 6:18 pm

foldy wrote:Don't get mad with bruce's pragmatic answers....

I'm not mad, and I don't find imagining some unknown, exorbitant expense to be "pragmatic" (unless you include Webster's "...to the exclusion of intellectual ... matters"). :P
User avatar
ComputerGenie
 
Posts: 235
Joined: Mon Dec 12, 2016 4:06 am

Re: intel Iris 6200

Postby davidcoton » Fri Jan 06, 2017 11:22 pm

ComputerGenie wrote:
foldy wrote:Don't get mad with bruce's pragmatic answers....

I'm not mad, and I don't find imagining some unknown, exorbitant expense to be "pragmatic" (unless you include Webster's "...to the exclusion of intellectual ... matters"). :P

Why do you consider development costs to be "unknown" and "imagined"? They are of course estimates, and notoriously difficult to predict, but as most software managers can tell you, estimates of the cost of software development are more often "under" than "over". Unless you want to make a very large donation to Pande Group to invest in software development, you just need to accept that development costs are professionally assessed, standard cost/benefit rules are applied, and the available development resource is applied in a cost-effective manner. (Usually, but not always, that will favour scientific productivity over donor experience.)
Image
davidcoton
 
Posts: 952
Joined: Wed Nov 05, 2008 3:19 pm
Location: Cambridge, UK

Previous

Return to New GPUs (whitelist)

Who is online

Users browsing this forum: Yandex [Bot] and 1 guest

cron