On Thu, Oct 18, 2012 at 12:31:56PM -0500, Gregory Carter wrote: > Thank you for clarifying the current issues. > > Slow is a relative term, I did not mean to say the project is slow > per se, just in comparison to Intel's roadmap the immediate benefit > to an open architecture shouldn't be the only reason for pursing the > project.
Intel pays a lot of salaries to ensure that a couple hundred (or thousand?) engineer-hours a day are spent ensuring the roadmap gets implemented. The fact that Open-Graphics has produced a board and a design based mostly on 2-3 people's work is pretty impressive, IMHO. > Bluntly I would like to see a hardware graphics architecture with > shader support that is not encumbered with patents or the typical > secret dealing requirements of signing over your first born in > meetings that happen at Nvidia and ATI to insure nobody knows how > much each company is stealing from their competitors or the open > source community. > > Which is the real reason for NDA's, and no source code. > > In your estimation, what would it take to build a prototype with > decent shader support for something like OpenGL 4.2 support? I'd like to see this too. I'm debating cashing out part of my 401k so I can spend a couple of months full time working on this. Would you be willing to put up some cash (or equivalent engineering time) to make this happen? I think this could be a great crowdfunding project, but its going to take lots of people to put in time and/or money. Like Timothy said, I'm probably going to do this regardless of if anyone else wants to contribute, and I'll probably start with modifying an electric resistance furnance to see if I can actually fabricate working silicon in a garage. But if I'm doing this myself, I expect I might be able to fabricate the past OGP1 design by about 2025. So here's my challenge to the community: If you want a 'faster' project, put some money or time in. > -gc > > > > > On 10/18/2012 11:46 AM, Timothy Normand Miller wrote: > >There are some people who might look at this "Agenda 21" thing and > >dismiss it outright as YACT (yet another conspiracy theory). > >Personally, I know nothing about it, so I can't comment on that. > > > >What's on-topic here is that 120 Watts is a challenge to dissipate, > >and you've only covered your CPU. Add to that another couple of CPU > >dies, and your main memory, and all your conversion losses, and then > >your GPU, and you've got a row of furnaces in your datacenter. > > > >This "power wall" we keep going on about is a real problem. The > >original Itanium (Merced) was limited in performance not by signal > >delay but by heat dissipation. The clock speed and voltage had to be > >lowered just to make it possible to cool the chip. Later Itaniums > >implemented more sophisticated power management that would dynamically > >adjust voltage and frequency to stay within power and temperature > >limits. Low-end (e.g. ARM) and high-end (e.g. Xeon) make active use > >of clock gating (to cut switching power) and power gating (to cut > >switching and leakage power) for idle circuits, and it's getting to > >the point where the granularity of power gating is going to have to be > >sub-processor in order to continue to scale performance. This is the > >"dark silicon" problem. Tons of interesting research is being done on > >low-voltage circuits as yet another approach to dealing with both > >power limits and to improve energy efficiency in general. > > > >As for "this" project being slow: With regard to the currently active > >OpenShader sub-project, the contributors have been myself and Mark > >Marshall. I put this out to the community in the hopes that we'd get > >some outside help and move the project along more quickly. But > >frankly, I'm going to work on this with or without your help, and it's > >the FOSS community's loss if they don't get any say in the project > >direction and if things get done a lot later than they would prefer. > >What I can do right now is get students involved, and I'm working on > >that, and I can apply for NSF and other grants, and I'm working on > >that too. But I have other pressures. I have two major research > >directions. One is energy efficiency in general (related to my prior > >research), which I'm working to apply to GPUs, and then there's this > >OpenShader sub-project, which is infrastructural. Ideally, I'd do > >OpenShader first because it would facilitate all other research in the > >area, but practicality may force us to compromise. The order and > >proportion to which I focus on each one is going to be a function of > >the help I can get, because I personally can only scale so far, and > >Mark works on this entirely in his spare time. There needs to be more > >than two people on this. > > > >What is the straw man here? The fact is, this infrastructural work is > >going to be of immense value to researchers in GPU-related areas. > >Whether or not you see it as feasible to turn this into consumer-grade > >hardware is another matter. But if I have a complete design on my > >desk (part of the whole process of back-annotating a simulator with > >real timing and energy information), I am in an excellent position to > >have this funded and fabricated. > > > > > >On Wed, Oct 17, 2012 at 8:42 PM, Gregory Carter <[email protected]> wrote: > >>Critical thinking? > >> > >>That is for process and materials selection for GPU design. > >> > >>Agenda 21 has nothing to do with critical thinking, it simply _is_. > >> > >>Good Heavens! > >> > >>I would like to state for the record that no where did I say that document > >>in my previous citation written by the "United Nations" contained _any_ > >>critical thinking. If you read it, it is entirely for a different audience. > >>(i.e. That audience I can assure you is not anyone on this list.). > >> > >>I am not from a defense contractor. > >> > >>I am simply interested in the idea of a open GPU. > >> > >>I think the larger question is exactly what Intel is going to do with its > >>roadmap. > >> > >>Did any of you guys look at this article: > >> > >>http://www.techradar.com/us/news/computing-components/graphics-cards/intel-could-kill-performance-pc-graphics-in-2015-1089339 > >> > >>Given how slow the project is moving, I am wondering if we are chasing a > >>straw man here. > >> > >>If the GPU simply becomes part of the CPU and just as programmable, the > >>issues with ATI and Nvidia might be a mute point. > >> > >>Which would be fine with me, ATI and Nvidia deserve to burn in my opinion > >>because I consider open source one of the few redeeming qualities of human > >>pursuits in science and technology and I must say, on a daily basis I have > >>to deal with these two companies and it physically makes me nauseous. > >> > >>-gc > >> > >> > >> > >> > >> > >> > >>On 10/17/2012 06:23 PM, Troy Benjegerdes wrote: > >>>Let's use some critical thinking here, okay? > >>> > >>>First, if a graphics card is using more than 60W power at *idle*, > >>>it's going to turn into a puddle of molten slag under full load. > >>> > >>>Second, start thinking about who stands to *make money* here. Who's > >>>going to get a bonus (or a promotion) because of some silly limit > >>>like this? > >>> > >>>The 'agenda21' thing sounds like someone from a defense contractor > >>>wrote it, so they can sell more military equipment. But the funny > >>>thing is we have a competitive world market, and if AMD (or an > >>>up-and-coming Chinese chipmaker) figures out theres more money in > >>>120W graphics cards, then that's what we'll have, and the EU can > >>>buy the slower limited versions and their players will get fragged > >>>more. > >>> > >>> > >>>On Tue, Oct 16, 2012 at 09:02:27AM -0500, Gregory Carter wrote: > >>>>Welcome to Agenda 21. > >>>> > >>>>http://www.un.org/esa/dsd/agenda21/ > >>>> > >>>>It is a well known publication by the U.N. to de-industrialize the west. > >>>> > >>>>It's main goals are: > >>>> > >>>>1) Stop or freeze/reduce the development of Urban centers. > >>>>2) Reduce the use of technology in various areas in human > >>>>communities to prevent the spread of electrical consumption and > >>>>preserve woodland areas and water reserves. > >>>>3) Reduce population through the application of genetically modified > >>>>foods, targeting high growth populations in Africa and other > >>>>undeveloped nations through the use of trade and GMO seed sales > >>>>controls. > >>>> > >>>>It not only means no GPU's in your future, it also means death for > >>>>billions. > >>>> > >>>>-gc > >>>> > >>>>On 10/15/2012 12:13 PM, Dieter BSD wrote: > >>>>>Why not limit power directly rather than limiting memory bandwidth? > >>>>> > >>>>>The US has a gas guzzler tax for cars that use a lot of fuel. > >>>>>But light bulbs and now GPUs get an outright ban? > >>>>> > >>>>> > >>>>>http://www.techpowerup.com/173706/New-EU-Energy-Guidelines-Could-Cripple-High-End-Graphics-Cards.html > >>>>> > >>>>>Hmmm, maybe the 1% think Big Bird is using too much energy? > >>>>>_______________________________________________ > >>>>>Open-graphics mailing list > >>>>>[email protected] > >>>>>http://lists.duskglow.com/mailman/listinfo/open-graphics > >>>>>List service provided by Duskglow Consulting, LLC (www.duskglow.com) > >>>>> > >>>>> > >>>>_______________________________________________ > >>>>Open-graphics mailing list > >>>>[email protected] > >>>>http://lists.duskglow.com/mailman/listinfo/open-graphics > >>>>List service provided by Duskglow Consulting, LLC (www.duskglow.com) > >>> > >>_______________________________________________ > >>Open-graphics mailing list > >>[email protected] > >>http://lists.duskglow.com/mailman/listinfo/open-graphics > >>List service provided by Duskglow Consulting, LLC (www.duskglow.com) > > > > > _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
