does anyone know what brand of monitor pairs best with nvidia 4090's? i know some monitors support nvidia while other support radeon better
Software G-Sync is pretty compatible, even with parts that haven't been tested/verified. I don't think I have any VRR displays that don't work well with my 4090 and only one of them is actually certified as G-Sync compatible.
So...what's going on here, have these two particular cards always been in short supply, or is it something to do with the 'new' RTX 5000 series being released soon?
I've got a suspicious mind that there's something being orchestrated somewhere ...but it's probably as I said, just me.
As I mentioned earlier, client GPUs are very low priority.
Every part comes as a certain die flavor that is then salvaged or crippled as needed to create a product with the capabilities required. As AI has taken off, client and enterprise needs have diverged to the extent that it makes sense to build specialized parts that cannot be turned into GPUs at all. Since this market produces several times the revenue of the gaming market, it takes priority. So, when NVIDIA negotiates their wafer allocation with a semiconductor fab, and decides what parts to built there, they don't order many top-end consumer GPUs. It's much more profitable for them to fill those AI orders.
Using the current/upcoming generation as an example: GB100 and GB202 are very different chips, but of similar size and manufacturing complexity/cost. GB100, which has sacrificed nearly all graphics capabilities for pure compute performance, goes into enterprise DGX accelerators that sell for mid-five-figures per GPU and are sold in huge quantities. GB202, which costs pretty much the same to make, goes into the RTX 5090 and the (as of yet unannounced) RTX 6000 Blackwell edition. These parts sell for the low four-figures (RTX 5090) to the mid-to-high four figures (RTX 6000).
NVIDIA can sell all the enterprise stuff they can have made at a ~1000-2000% markup. Even the most expensive client parts only sell at 100-300% markup.
It was the same last generation, though to a lesser degree. Since these orders need to get to fabs months or years in advance, they have to try to estimate the relative demand of each market segment. As the AI boom has been going on for years and shows no signs of abating, they've settled into allocations that are very enterprise heavy, because that's where the money is. This means a shortage of high-end client GPUs at any price, because any quantity they've seen fit to have made will sell out at almost any price.
The RTX 4090 was never a high volume part. It got the binning rejects from the RTX 6000 Ada edition, which was itself relatively low-volume, and mostly existed as a combination of marketing exercise and way to salvage silicon that couldn't be used elsewhere. The low stock of last gen parts also further increases demand for the new generation as varonica mentions.
This is what annoys me about Big Business. As if their profit margins aren't big enough already.
A wise human once said "money is the root of all evil". The older I get and as the years go by, it seems this adage grows ever more accurate.
Money, by it's very nature, rewards scarcity and inequity. Mediums of exchange and stores of value have no utilitity if no one wants for anything.
As for big business, policy changes in much of the world, especially since the 1970s, have increasingly conflated it with government. The most efficent way to extract profit is with coercive monopolies that are tacitly supported by nation states. Government of corporations, by corporations, for corporations...all wrapped in a sweet candy shell of democracy.
The biggest thing separating reality from the fantasy dystopias in games like Elite and Cyberpunk is that reality is just a little less on the nose and our villians have better publicists. Oh, and that even railroady games tend to give us more opportunity and agency than life does.
