NVIDIA GeForce RTX 5090 Founders Edition Unboxing

[LittleBro]

Are you implying that DLSS is unnecessary?

Exactly, it is NOT something we can't live or game without.

 

[Delta6326]

@W1zzard are you able to give the length of the adapter cable provided? Just curious if it's long enough so I can hide the four connectors for cable management in my custom wood case.

 

[usiname]

I am mostly curious about the full load power consumption and the temperatures with the new 2-slot cooler.
If the power consumption actually hit around 600W we will have new wave of melted connectors

 

[Vya Domus]

Yeah, but adds unecessary cost

This thing is 2000 dollars man.

 

[mtosev]

I'm waiting for the reviews to come out. And oh the prices here in Europe.

 

[JustBenching]

This slide is an incredible scam, why on earth at 35fps native resolution would I have 140ms of latency and not 30ms?
My god Nvidia, my goodness...

View attachment 380723

Because latency isn't entirely dictated by your framerate. CPU prepares a frame - GPU is still processing the previous frame -- framerate is 35 fps but input latency is still much higher than 30 ms because the frame was prepared 100 ms in the past. Simple, right?

 

[csendesmark]

Still under NDA as per page 4.

Is 30th of January the date?

This always makes me chuckle

 

[Bomby569]

Is 30th of January the date?

that's under NDA too. You can only see the cardboard box. Enjoy your free teaser

 

[Icon Charlie]

As I have stated before, I've done manufacturing, worked in manufactoring, as well as reporting on manufactoring companies. So you understand where I am coming from.

I'm going to call part of the article/press release from Ngreedia is BULLsh!t Marketing.

If I remember correctly by law here in the US You have to use Bio-degradable inks for boxing.
The glaze look on a fancy box is most likely clay based and not plastic based as it has to be biodegradable for landfill here in the US. As far as plastic wrapping for cables/other components? I haven't seen that for years.

Maybe they are saving the earth by not giving me a rubber band that may tie up the other components in the box.

OR maybe they are saving money by not giving me that rubber ban that may tie up the other components that might be in the box.

I don't like BULLsh!t Marketing.

For Ngreedia, they are cutting every corner to make money off of the consumer. Their Eco Marketing comments = Treehugger coomby ya feel good moments. This is nothing but going Ngreedia on the cheap to make additional profit per unit and give a free "feel good monent in the process".

And saying out of the goodness of their decayed hearts that they are saving the planet by going "eco"???.

Yea... sure... Go con someone else with your marketing drivel. You did it for the $$ signs.

If I am going to pay over $2000 dollars for a (used car) !@#$ gaming card, it better be in a nice video card box so I can show the console peasents the awesomeness of the purchase.

IF AMD pulled the same sh!t on me or anyone else trying to play the Eco card on me like this. I call them out as well.

In the end it is all about the money. Nothing more. Nothing less.

ADDED: I WAS going to buy something for myself as I've done well this year despite all of the @$$ hattery going on with the economy. But man this set me off.

Instead I think I'll buy another piece of investment property and maybe in 6 months I'll look into upgrading my rig. Sometimes when you don't get your drug of choice (coffee for me) and you read this type of nonsense from a monopolistic corporation, it kinds of sets your mood for the rest of your day.

 

[MxPhenom 216]

Is 30th of January the date?

This always makes me chuckle
View attachment 380744

24th is when the review NDA is up on the 5090. Market availability the 30th

 

[dyonoctis]

I see they copied even the cheap packaging from Apple.

The only adjective that I would use for an apple packaging is "small" but cheap? The Mac studio that I'm using at work came in a box that's intricate as f*** with a canvas handle and some kind of cushion system made of folded cardboard at the base, base wich seems to be molded. The thick hard cardboard that Nvidia used before and that Apple is using on their other product, is annoying IMHO, the cover doesn't always slide very well. Especially when you keep them in storage, and there's not enough weight inside anymore

 

[evernessince]

It's amazing how many people will rush out to buy this at $2000 so they can get a few more frames while playing on, most likely, a 1440 or 1080p monitor. Psychological obsolescence lol. The same people will scream about housing unaffordability and car insurance prices.

Yeah, that's called revenge spending. People can't afford the actual milestones in their lives anymore (housing, kids) so they splurge on luxuries they can afford. It's really common nowadays and it's a symptom of the times.

EDIT**

And before anyone else feels the need to virtue signal, saying I'm criticizing people for their purchases, the above is not judging anyone. Revenge spending is an observed psychological phenomenon. and pointing it out isn't me being judgemental.

 

[Bomby569]

everybody is discussing the cardboard box, Nvidia has already won

Yeah, that's called revenge spending. People can't afford the actual milestones in their lives anymore (housing, kids) so they splurge on luxuries they can afford. It's really common nowadays and it's a symptom of the times.

A fool and his money are easily parted

 

[the54thvoid]

Yeah, that's called revenge spending. People can't afford the actual milestones in their lives anymore (housing, kids) so they splurge on luxuries they can afford. It's really common nowadays and it's a symptom of the times.

That's not really a logical argument. There are also plenty of people who can afford such luxuries and are willing to pay for them. Chastising people on the forums for their buying choices is never a good platform to pitch your argument. The same applies to those who 'poverty' shame others for not paying more to get a better card.

Is the card worth the money? Not by a functional metric, but Nvidia pitch it as a halo product. It's just how it is. And blaming individuals for purchasing it is futile, especially when they get bought by businesses en masse.

 

[MxPhenom 216]

It's amazing how many people will rush out to buy this at $2000 so they can get a few more frames while playing on, most likely, a 1440 or 1080p monitor. Psychological obsolescence lol. The same people will scream about housing unaffordability and car insurance prices.

I dont think a majority of folks planning to buy a 5090 is doing so for gaming.

 

[Solid State Brain]

I've been considering a 5090 for LLMs but for the price it should have had at least 48GB of VRAM. The Founder's Edition 2-slot size even makes it suitable for multi-GPU configurations, but on the other hand the TDP and the ridiculous amount of PCIe cables required appear to have been placed to discourage such uses.

 

[Bomby569]

I dont think a majority of folks planning to buy a 5090 is doing so for gaming.

majority! sure that would never happen, but if you look at the Steam survey and count only the last gen desktop cards, percentage wise, there is a lot of people using the 4090 to game, more than one would reasonable assume would ever happen. I don't see any reason why that should change.

 

[LittleBro]

Because latency isn't entirely dictated by your framerate. CPU prepares a frame - GPU is still processing the previous frame -- framerate is 35 fps but input latency is still much higher than 30 ms because the frame was prepared 100 ms in the past. Simple, right?

I'm sorry, I don't want to be rude but this is utter nonsense what you wrote.
CPU computational part was surely not done 100ms before GPU when fps is 35.

In scenario with 35fps, time between frames is around 28.5ms. 100ms would on CPU side cause GPU to become bottlenecked by CPU. Having 35fps, this is most probably the opposite situation - bottleneck is on the side of a GPU. As a kind of remedy to this situation, DLSS renders in reduced resolution, which allows for improved ramerates, thus frame time decreases and that lowers latency, but only so long as CPU does not become a bottleneck. No such thing as waiting 100ms for GPU to render frame exists with 35fps.

Nvidia was trying to point out overall system latency: CPU latency, GPU latency, monitor latency, input devices latency. You can have half overall system latency than what Nvidia showed even with 35fps.

The only adjective that I would use for an apple packaging is "small" but cheap? The Mac studio that I'm using at work came in a box that's intricate as f*** with a canvas handle and some kind of cushion system made of folded cardboard at the base, base wich seems to be molded. The thick hard cardboard that Nvidia used before and that Apple is using on their other product, is annoying IMHO, the cover doesn't always slide very well. Especially when you keep them in storage, and there's not enough weight inside anymore

That's like $10-15 costs packaging only. What an overkill. Such a waste of resources.

 

[JustBenching]

I'm sorry, I don't want to be rude but this is utter nonsense what you wrote.

CPU computational part was surely not done 100ms before GPU when fps is 35.

In scenario with 35fps, time between frames is around 28.5ms. 100ms would on CPU side cause GPU to become bottlenecked by CPU. Having 35fps, this is most probably the opposite situation - bottleneck is on the side of a GPU. As a kind of remedy to this situation, DLSS renders in reduced resolution, which allows for improved ramerates, thus frame time decreases and that lowers latency, but only so long as CPU does not become a bottleneck. No such thing as waiting 100ms for GPU to render frame exists with 35fps. It may theoretically become a bit close to reality with 10 fps, assuming that GPU computational latency is zero, which is not, so it would be 9.xx fps.

Nvidia was trying to point out overall system latency: CPU latency, GPU latency, monitor latency, input devices latency. You can have half overall system latency than what Nvidia showed even with 35fps.

If you don't know how it works you don't know how it works. You can read about it before calling it utter nonsense. By default - and obviously depends on the game / engine, CPU prepares frames in advance. 2-3-4 of them. If your framerate is at 30 and GPU bound, your GPU is still processing frame 1 while CPU has frame 4 ready and waiting. So add 30ms on each one of them and you get to over 100 ms latency.

 

[matar]

I am happy with what i have and not upgrading AT ALL if the RTX 5070Ti was $499 then maybe or the RTX 5080 was $699 then ok Nvidia i may upgrade what happened in 2018 GTX 1080TI flagship was $699

 

[W1zzard]

@W1zzard are you able to give the length of the adapter cable provided? Just curious if it's long enough so I can hide the four connectors for cable management in my custom wood case.

https://imgur.com/zTREGr9

 

[MxPhenom 216]

majority! sure that would never happen, but if you look at the Steam survey and count only the last gen desktop cards, percentage wise, there is a lot of people using the 4090 to game, more than one would reasonable assume would ever happen. I don't see any reason why that should change.

Steam survery does not tell you anything about users outside the gaming sphere.

Within my circle, i know 1 person with 4090 for gaming. I know 20+ others who have them for CAD, Ai/LLM stuff. Small sample size, but businesses flock to these cards for workloads that arent gaming, especially if they dont benefit at all from Quadro.

 

[LittleBro]

If you don't know how it works you don't know how it works. You can read about it before calling it utter nonsense. By default - and obviously depends on the game / engine, CPU prepares frames in advance. 2-3-4 of them. If your framerate is at 30 and GPU bound, your GPU is still processing frame 1 while CPU has frame 4 ready and waiting. So add 30ms on each one of them and you get to over 100 ms latency.

Dude, if what you state is true, than you could see impact of your movement with mouse only each 2-3-4th frame.

There is a rendering and computational queue but it does not work like how you described. CPU as well as GPU does continous computing, nothing like 3 frames and 100ms waiting. Some computational results, mostly those which does not change with every frame, can be queued to be rendered later (depends on frequency of particular effect showing up in the game - e.g. weather clouds position change, sun position change).

 

[JustBenching]

Dude, if what you state is true, than you could see impact of your movement with mouse only each 2-3-4th frame.

No...?

 

[Denver]

I see they copied even the cheap packaging from Apple.

Let's cut a few corners on the $2000 product, mask it with an "environmentally friendly" label, even though the GPU's TDP alone is nearly 600W. - Jensen
I don't see how it wouldn't work.

I know this might not be the right topic—or maybe there isn’t one—but have RDNA4 samples actually arrived for review, or are they still just an urban legend? @W1zzard