It’s no secret that higher pixel counts are harder to achieve since computational power required grows approximately quadratically while GPU and CPU power improvements are more linear.
A lot of laptops still don't support HDMI 2.1, and DP 2.0/2.1 only just came out (and I don't know if any displays are using it yet). There's also the bullshit the HDMI forum pulled where they relabeled HDMI 2.0 as HDMI 2.1 (years after HDMI 2.1 was out and established as "the one that can do 4k120hz no chroma subsampling" (a lot of implementations were around 40Gbps instead of the full 48Gbps because it was cheaper, and they didn't need the full 48Gbps for 4k120hz 10bit)), despite them having hugely different specs.
Displayport 2.0/2.1 is also a clusterfuck because they decided that it would max out at either 40Gbps or 80Gbps depending on the implementation (DP40 and DP80), and companies don't generally tell you what their implementation actually supports.
Macbook Pros only got HDMI 2.1 this year, iirc base and airs still don't have it. Intel iGPUs only do HDMI 2.0 and DP 1.4, AMD's Zen4 laptops can do HDMI 2.1 and DP 2.1 (idk what bandwidth). For dGPUs, AMD's 6000 series is HDMI 2.1+DP 1.4 (but there are issues with DSC), while 7000 series is HDMI 2.1 and DP 2.1 (DP80). Intel Arc dGPUs are DP 2.0 (DP40) and HDMI 2.1. Nvidia's 30 and 40 series are HDMI 2.1 and only DP 1.4.
edit: Also, people don't realize just how fucking huge an 8k framebuffer is, and you need multiple for a swapchain. Even for just basic 8bit colour you're looking at ~133MB per framebuffer.
It's for large monitors. I run a 43" 4k tv as a monitor. I would absolutely benefit from a similar sized tv in 8k. And people who run closer to the 50" size would benefit even more.
I think you're probably right about traditionally sized monitors, though.
I’m a bit surprised we don’t see point-to-point ethernet cables doing this. They are standardised, reliable, and carry a lot of data. My napkin math says an 8K display at 40 bits per pixel and 60 fps is about 18 Gbps, well into the range of fancy ethernet.
And, if consumer electronics can drive down the costs of these interfaces, everyone wins.
Anyway, with compression it should be a piece of cake. I never quite understood why HDMI even streams pixels and not refresh instructions.
That's not a lot for people in the market for 8K monitors. 24G VRAM in the form of used 3090 cards is way cheaper than high end monitors. But yeah lack of DP2.1 support on current-gen cards like the 4090 is a problem
> you need multiple for a swapchain
There should be just two with variable refresh rate
People want 8k for better PPI for text rendering, not for games, meaning they want to use it with their work laptop.
> There should be just two with variable refresh rate
There's still a lot of software out there that's using DXGI blit model instead of DXGI flip model, and a triple buffered vsync+windowed blit app means you're looking at 5? framebuffers in flight at once. (ignoring any additional intermediate buffers for each app, and potentially having multiple overlapping windows drawing onscreen at one time).
Thanks to Moore's Law, computer power has been growing exponentially since forever. Single-core performance has stagnated, sure, but total performance is still following the curve. Graphics is easy to parallelise, so throwing more cores at it is not a problem. In fact, this is precisely what GPUs do.
In other words, the available "operations per pixel per second" have gone up over the years, despite the increasing resolution!
In fact, we're at the point now that we can even do real time ray tracing, which was unthinkable at any typical resolution just a decade ago. Top-end RTX 4090 GPUs can even do ray tracing at 4K, which is just crazy.
Nitpick: Moore's law is about number of transistors on microchips not performance.
The original general-purpose GPU, the NVIDIA 8800 GT had 112 cores. The RTX 4090 has 9,728! Each core has pretty much remained the same: a 32-bit arithmetic/logic unit (ALU).
Vertical scaling is of course possible, but it’s not really useful for driving a higher resolution monitor in your living room because of bottlenecks being completely elsewhere.
So are you saying that we should go back to "just wait for the picture to render..." era just because we can have higher density?
Top-End RTX 4090 cater for a niche market.
Data from sensor and to display are usually (but not always) raw and uncompressed.
To go beyond 8k 120fps we will need fiber, or better compression.
We have good codecs to move data online. And specialized ASIC to do live encoding with good quality.
The current compression defined by HDMI is very rudimentary.
Quite ironic to complain about waste at the same time as making this pointless and aggressive comment
Maybe you don't realize you're on a worldwide forum where anyone can respond to anyone else for any reason? I'm perfectly entitled to respond to your comment as someone who does not game. The person you replied to isn't the only person in this entire thread. And suggesting that it's strange for me to reply to your comment is just strange in itself.
Do they?