At 02:35 PM 4/14/2007, Brad A. Steffler wrote: >Frank, > > >I think most amateurs fiercely resent the DRM intrusiveness of >Microsoft's Vista. I know I do.The future does appear very nice for us >freedom >loving hams. Just yesterday I read that the frame buffer(s) in AMD's new >GPU will not be accessible outside the GPU. >( http://www.infoworld.com/article/07/03/28/14OPcurve_1.html ). This is, >apparently, a device to placate the "premium content providers" >(read RIAA and MPIAA and their member businesses). It also destroys one >reason why we buy PC's in the first place - freedom of flexibility.
I can think of several non-evil reasons why not to make the frame buffer accessible outside the GPU.. I'm doing a design right now using an FPGA that takes compressed video in a stream of IP packets and ultimately has to generate composite analog video. I'm hanging some RAM off the FPGA to serve as a frame buffer, but that RAM is ONLY connected to the FPGA. The decompressor writes and the video output player reads from the buffers and spits out to a DAC. Sure, I could implement access to the RAM from the processor bus, but it would be a pain, because I'd have to figure out a way to arbitrate between bus and decompressor and video output. Arbitrating between 3 address and data sources is not as easy as two, especially if I have to do it in an interleaved way (since the video stream is, by definition, nice and interleaved in time..). Sure it *can* be done, but I don't need to, so I won't, except perhaps in some hideously hacked up debug form (e.g. running a data bus out to some pins for logic analyzer access) that costs almost nothing, and has no performance impact. Chip space and design time is expensive. If there's no need for direct access to the frame buffer, then why provide it. It has to be designed, tested, drivers written, etc., and it just provides more opportunities for bugs to creep in. And, in the commercial world, it would just potentially extend the "time to market", which is a form of business death. Consumer video interfaces are a VERY cost competitive market with very small margins to pay for added functionality. The only reason you'd provide direct buffer access was to retain some backward compatibility, and even then, at some point, you just say: emulate it in software... Even as long ago as the VGA era, backwards compatibility was a design curse, just to keep some software that did direct buffer writes working. _______________________________________________ FlexRadio mailing list [EMAIL PROTECTED] http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/ FlexRadio Knowledge Base: http://kb.flex-radio.com/ FlexRadio Homepage: http://www.flex-radio.com/
