They'd still sell both, even if they both were proprietary, in much the same way that motherboard manufacturers tend to support both AMD and Intel CPUs. Not on the same board, no, but as part of two product lines from the same company.
unlikely.
right now gsynch's problem isnt technology but economics.
gsynch exists as a value add attempt by nvidia that certainly answers a problem in the gaming market. however because there was no standard on variable refresh timings/protocols and because of some sort of limitation on the gpu side, it is necessary to have a custom panel controller. they are currently using a fpga on a custom daughterboard shoehorned into a few 120 hz tn models. this would normally be done with an asic fixed function chip that is ordinarily an off-the-shelf kind of component.
but since lcd controller chip making is a niche industry there are probably only a handful of makers who dont update their designs until there is a major change in standards or must-have features. so nvidia has to use the more expensive and somewhat wasteful fpga. this is the bulk of the cost of the gsynch premium , though the insurance coverage for the housecall/mail-in service to install the daughterboard probably adds into it.
if there was an asic gsynch controller the cost would probably only be a few dollars over the normal BOM. but designing and contracting out a fab to make such a narrow use chip is very expensive and usually amortized over years and mass volume. for nvidia to commit the millions necessary to make a gsynch asic, they would need massive contracts with the bulk of lcd oems. given that pure gaming monitors make up a smaller portion of all lcd models, that is a difficult sell.
it is a chicken or the egg conundrum:
until there is a clear demand for variable refresh monitors, not enough oems will commit to a long term big volume contract for gsynch asic controllers. if nvidia cant get enough contracts for the controllers, they cant justify spending the millions necessary to make an asic. without sample chips no oem can put out a test model to check the waters. this is why nvidia went the expensive fpga route.
if nvidia had the market to itself this is a fine and admirable effort to push technology, but amd looked back at its ip portfolio and found some of their early efforts in variable refresh. the fact that they already had a version of variable timing built into their gpu design for the laptop panel eDisplayport interface was probably serendipity.
because the protocol for freesynch is so closely tied to displayport, it is an easier proposition to amend the vesa standard and have the niche controller asic makers enable the dp functionality in all their chips for the next iteration. since it is open it means intel can use the asynch function as well.
monitor oems are more likely to drag their feet and wait for an industry standard and protocol built into an off the shelf component like the controller chip, than they are to commit all their product lines to a proprietary gsynch chip that in the near term adds a $200 premium.
gsynch is a polished and finished product, but that isnt enough to make it a certain thing. No amount of gamer enthusiasm will change the economics of monitor oems.
Facebook
Twitter