Date: Tue, 27 Oct 1998 21:07:28 -0500 From: James Higgins <higginsj@iname.com> To: questions@FreeBSD.ORG Subject: [Fwd: Re: I want a 3D accelerated X-se[r]ver!] - From Nvidia About the Riva TNT Message-ID: <36367C60.8CD6EA66@iname.com>
next in thread | raw e-mail | index | archive | help
Hello, I saw a question about this and I sent the guy to info@nvidia.com. Looks like they have been getting alot of response. I kind of understand the viewpoint of nvidia. I still want OpenGL acceleration though. : ) Thought I would pass it along for all interested.... James Higgins -------- Original Message -------- Subject: Re: I want a 3D accelerated X-se[r]ver! Date: Tue, 27 Oct 1998 17:06:05 -0800 (PST) From: Jim Putnam <putnam@NVidia.COM> To: John.McKenzie@delta-air.com, Rickard.Eneqvist@swipnet.se,Stefan.Pedersen@ks.ericsson.se, UNT012@ibm.rhrz.uni-bonn.de,aod@unisys.com.br, bg19056@binghamton.edu, bg19739@binghamton.edu,brion@pobox.com, budnik@alpha.rada.kiev.ua, cccomputers@iname.com,cfriesen@home.com, cicotti@cli.di.unipi.it, dimitri@ibb.net,doctor.stein@usa.net, echampion@home.com, fredrik@reite.com,gus@e-chemie.tu-darmstadt.de, hans.jones@midatorer.se, hemenk@infi.net,higginsj@iname.com, ian@greetme.com, irsman@iag.net,jarlaxle@falcon.cc.ukans.edu, jason@alltel.net, joerg@suse.de,jon.davis@lmco.com, jtew@ihug.co.nz, karsten.hahn@stud.uni-karlsruhe.de,kmoore@frognet.net, kovacsp@egr.uri.edu, krauss@medizin.fu-berlin.de,kress@datacomm.ch, locogris@arnet.com.ar, martin.gaebler@TalkNet.de,mavrick@sinesurf.co.nz, mbenson@magideas.com, mechanix@digibel.org,mholling@coginst.uwf.edu, moooster@uclink4.berkeley.edu,morrill@austin.ibm.com, naeem@cyweb.com, nbecker@fred.net,ober37@mediaone.net, patrick@narkinsky.ml.org, pc@shaw.wave.ca,ralf.horns@itzehoe.netsurf.de, reid@lanl.gov, rjd3@bigfoot.com,rob@austin.rr.com, rw@dialup.ptt.ru, sasa.mihajlovic@hpt.hr,simba@leontief.ru, sjoel@sjoel.xs4all.nl, souser@gis.net,souser@windoms.sitek.net, tnaleid@uswest.net, tozwierz@medianet.pl,wlj@cais.com, woland@cbr.astrakhan.su, wolfrum@cs.uni-bonn.de,xray@friko.onet.pl Thank you (all) for writing. Since there seem to be quite a lot of you all concerned about the same topic, I thought I'd take the liberty of answering you en masse. If I didn't answer your question in this mail, please feel free to contact me individually. I'm sorry for the length of what is to follow, but the issues you've raised are complex. First, for those of you happy with the purchase of your RIVA accelerator, thank you. We work hard to provide world-class graphics at an affordable price, and we realize you have options when you make such a purchase. For those of you who are unhappy because you do not have Linux support when you thought you did or should have, we're genuinely sorry. It was never our intention to mislead anyone as to the kinds and levels of software support available for our products. As you may or may not know, NVIDIA is in the business of providing graphics accelerator chip designs to chip fabrication firms and large OEMs, and indirectly, to system integrators. Many of our OEMs also produce products for the retail market. We don't market graphics boards ourselves, nor do we have any fab capacity; those are done by our customers. As with any operational business, we listen very carefully to the needs of our customers. Unless we meet those needs, we go out of business. To date, the requirements of our customers have been unambiguous; their primary emphasis is high 2D/3D performance on the Windows operating systems through the implementation and support of standard 2D and 3D programming interfaces (which means OpenGL and D3D). We do the Windows drivers and library implementations ourselves because we know the chip architecture best. Many of our OEMs take our drivers and modify them because they have favorite control panels and things, and some of them add functionality to our chip (no two board manufacturers do video in quite the same way, for example), and support that through additions to our drivers. We have not had a customer require Linux support, it seems that the majority of their customers don't demand it. It would be nice if this would change. Informing your favorite board vendor of your requirements would help. One of the things we decided some time ago was that we would not be in the business of trying to write anything other than the Windows code ourselves. Linux is by no means unique in this respect, we also don't provide any direct support for OS/2, BeOS, Solaris, or any of the other fifty or so potentially interesting operating systems out there. This is partly because our customers haven't required it, and partly because we must devote all of our effort to refining and improving the things our customers do require. This will not change soon if at all; we are not going to begin doing drivers for random operating systems ourselves. All of our customers are under heavy non-disclosure agreements. They also get a considerable amount of support from us in the form of source code, and assistance for answering questions they have. As such, we have not developed documentation for our chipset which is adequate to be used without some form of support. This is not by design, it is a consequence of six month design cycles, and tight deadlines in the consumer marketplace. A chip specification for a design as complex as ours without support is useless, as we've found out in the past. A second problem is that our chip specifications are the basis on which our company builds products, and as such is intellectual property. Access to those specifications without NDA protection or completed patent filings on all topics of interest violates the requirement to keep proprietary information private. So, we're in kind of a bind. We can't just hand out the bare chip specs to a company like Xi Graphics even under NDA (though they keep asking for them, and they don't believe me when I tell them this) without considerable support, which we can't afford to give them. We know you want accelerated graphics under Linux. The BeOS people want it just as bad, and the Solaris people do too. The list goes on and on. We have a program underway now to enable those people through information we'll make freely available on our website. This will require no NDA protection, what people do with their code is up to them. I can't tell you when the program will be ready, because I'm really bad at projections, but we hope soon. That's pretty much all there is to it. It won't be the answer many of you wanted, but I hope you'll understand our position, and further understand that we are aware of the problem, and are trying to find ways to fix it. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?36367C60.8CD6EA66>