The most interesting aspect here might be the improved tensor cores for AI workloads - could finally make local LLM inference practical for developers without requiring multiple GPUs.
Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
651–660 of 776 posts
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#652Earlier quoted context omitted.
https://www.youtube.com/watch?v=cXpoJlB8Zfg https://www.youtube.com/watch?v=iyn2NeA6hI0 Can you please point at the mentioned effects here? Immersion in what? Looks like PS4-gen Tomb Raider to me, honestly. All these water reflections existed long before RTX, it didn't introduce reflective surfaces. What it did introduce is dynamic reflections/ambience, which are a very specific thing to be found in the videos above.…
> by not implementing any RTX-off graphics at all. Just taking this one, you actually make a point about having a raytracing-ready graphics card for me. If all the games are doing the hard and mathematically taxing reflection and light-bouncing work through raytracing now and without even an option for non-raytraced, then raytracing is where we're going and having a good RT card is, now or very soon, a requirement.
It’s not me making this point, but nvidia’s green paper agreements with particular studios to milk you for more money for basically same graphics we had at TR:ROTT. If you’re fine with that, godspeed. But “we” are not going anywhere RT “now”. Most of Steam plays on xx60 and equivalents, which cannot reasonably run RT-only, so there’s no natural incentive to go there.
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#653Earlier quoted context omitted.
Rasterizing results in better graphics quality than DLSS if compute is not a limiting factor. They are trying to do an apples to oranges comparison by comparing the FPS of standard rendering to upscaled images. I use DLSS type tech, but you lose a lot of fine details with it. Far away text looks blurry, textures aren’t as rich, and lines between individual models lose their sharpness. Also, if you’re spending $2000 f…
DLSS is becoming the standard rendering.
It's not. It's becoming the standard lazy choice for devs though.
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#654Similar CUDA core counts for most SKUs compared to last gen (except in the 5090 vs. 4090 comparison). Similar clock speeds compared to the 40-series. The 5090 just has way more CUDA cores and uses proportionally more power compared to the 4090, when going by CUDA core comparisons and clock speed alone. All of the "massive gains" were comparing DLSS and other optimization strategies to standard hardware rendering. Som…
Jensen did say that in the presentation that compute performance isn't increasing at large enough scales to make enough change. The shift is moving to reliance on using AI to improve performance and there are additions in hardware to accommodate that. Isn't not being kept a secret, its being openly discussed that they need to leverage AI for better gaming performance. If you can use AI to go from 40fps to 120fps with…
DLSS and DLAA are terrible for any high-movement Games Like FPS, racing games, Action Games. I wouldn't exactly call it near identical quality. To shareholders this may ring true, but most gamers know that these GPS gains are not worth it and don't use it.. (They still buy it tho)
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#655Anyone has any info on Node? Can't find anything online. Seems to be 4nm but performance suggest otherwise. Hopefully someone do a deep dive soon.
TSMC 4NP process
Source: https://www.nvidia.com/en-us/data-center/technologies/blackw...
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#656Similar CUDA core counts for most SKUs compared to last gen (except in the 5090 vs. 4090 comparison). Similar clock speeds compared to the 40-series. The 5090 just has way more CUDA cores and uses proportionally more power compared to the 4090, when going by CUDA core comparisons and clock speed alone. All of the "massive gains" were comparing DLSS and other optimization strategies to standard hardware rendering. Som…
Flops went up 26% and power draw 28%.
So the biggest benefit is PCIe 5 and the faster/more memory (credit going to Micron).
This is one of the worst generational upgrades. They’re doing it to keep profits in the data center business.
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#657interesting launch but vague in its own way like the one from AMD (less so but in a different way).
it is easy to be carried away with vram size, but keeping in mind that most people with apple silicon (who can enjoy several times more memory) are stuck at inference, while training performance is off the charts through cuda hardware.
the jury is yet to be out on actual ai training performance, but i bet 4090, if sold at 1k or below, would be better value than lower tier 50 series. the "ai tops" of the 50 series is only impressive for the top model, while the rest are either similar or with lower memory bandwidth despite the newer architecture.
i think by now the training is best left on the cloud and overall i'd be happy rather owning a 5070 ti at this rate.
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#658I have a serious question about the term "AI TOPS". I find many conflicting definitions while others say nothing. A meaningful metric should at least be well defined on its own term, like in "TOPS" or expanded "Tera Operations Per Second", what operation it will measure? Seemingly NVIDIA is just playing number games, like wow 3352 is a huge leap compared to 1321 right? But how does it really help us in LLMs, diffusio…
The 5090 TOPS number is with sparsity at 4bits, so it doubles the value compared to the 8bit sparse number for 4090.
The real jump is 26%, at 28% higher power draw and 25% higher price.
A dud indeed.
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#659I have a serious question about the term "AI TOPS". I find many conflicting definitions while others say nothing. A meaningful metric should at least be well defined on its own term, like in "TOPS" or expanded "Tera Operations Per Second", what operation it will measure? Seemingly NVIDIA is just playing number games, like wow 3352 is a huge leap compared to 1321 right? But how does it really help us in LLMs, diffusio…
It would be cool if something like vast.ai's "DLPerf" would become popular enough for the hardware producers to start using it too. > DLPerf (Deep Learning Performance) - is our own scoring function. It is an approximate estimate of performance for typical deep learning tasks. Currently, DLPerf predicts performance well in terms of iters/second for a few common tasks such as training ResNet50 CNNs. For example, on th…
We don’t need this. We can easily unpack Nvidia’s marketing bullshit.
5090 is 26% higher flops than 4090, at 28% higher power draw, and 25% higher price.
Re: Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs
#660I have a serious question about the term "AI TOPS". I find many conflicting definitions while others say nothing. A meaningful metric should at least be well defined on its own term, like in "TOPS" or expanded "Tera Operations Per Second", what operation it will measure? Seemingly NVIDIA is just playing number games, like wow 3352 is a huge leap compared to 1321 right? But how does it really help us in LLMs, diffusio…
The 5090 TOPS number is with sparsity at 4bits, so it doubles the value compared to the 8bit sparse number for 4090. The real jump is 26%, at 28% higher power draw and 25% higher price. A dud indeed.
It really sucks. BTW, how did you find the statement? I cannot find it in any place.