News Nvidia Reportedly Readies 800W RTX 4090 Ti With 18,176 Cores, 48GB GDDR6X

[Admin]

Possible specifications of the GeForce RTX 4090 Ti have been leaked. Dubbed "The Beast" by Twitter's Kopite7Kimi, the GPU reportedly has 18,176 shader cores, 48GB of GDDR6X memory, and an 800W TBP.

Nvidia Reportedly Readies 800W RTX 4090 Ti With 18,176 Cores, 48GB GDDR6X : Read more

 

[peachpuff]

These should be bundled with a psu.

 

[Giroro]

I could write something snide about the 4090 Ti popping circuit breakers, but realistically those cards only exist to give away to YouTubers in the "have over 10 people on staff category". All of those people are running commercial power, so it really doesn't matter.

I'm a lot more upset about the idea their "4070" could be a 300W card. That's significantly higher than even a 2080Ti. I live in an area with this rare event called "Summer", which happens on a semi-regular basis. . .

A sub-200 Watt card might not even happen from Nvidia next generation, which is wild to me. If these are the results Nvidia is getting, then maybe they should finally remove the useless Ray Tracing software. Game devs don't want to support it. It's time to move on. Stop adding all that cost and power to gaming cards for tech that is really only useful to Hollywood render farms.

 

[Roland Of Gilead]

Wow, this is getting insane!

On average a 80-100w increase on most cards, with the 3090ti to 4090ti even more!

A 650w PSU with midrange Ada cards may not be sufficient! Lotta folks will have to consider better PSU's to upgrade.

 

[Giroro]

These should be bundled with a psu.

Don't worry, it will be.
But it probably will be a severely underspecced PSU, which might catch fire.

 

[peachpuff]

Don't worry, it will be.
But it probably will be a severely underspecced PSU, which might catch fire.

Gigabyte psu? 😂

 

[King_V]

Just wait 'til the RTX 5090 Ti comes out....

 

[Phaaze88]

This... this is really going to be a BFGeforce and not a Titan? I mean, 48GBs of Vram...

 

[husker]

Seems to be more for just bragging rights than a practical commercial product and to see how far they can push the limits of the technology. Kind of reminds me of a concept car, but in this case they can sell a few production units as well.

 

[spongiemaster]

Everything about this "leak" sounds like Nvidia just trolling people or trying to figure out who is leaking stuff on twitter. If the 4090 has a 450W TDP, how does the 4090Ti have an 800W TDP? Even Nvidia's upcoming enterprise H100 SXM boards top out at 700W. Based on specs alone, there should not be a 350W difference. I can't see Nvidia releasing a 48GB RTX card either. If this card exists in any form, it will be back to the Titan branding.

 

[10tacle]

Interesting specs of the 4080 and 4090 between CUDA cores and memory bus compared to the 3080 and 3090 (if this 4-series info is true) and then seeing where the 3080 TI is. I'm still scratching my head why Nvidia even created the 12GB version of the 3080 which just made it effectively a 3080 Ti at the same price point (and I'm a 3080 Ti owner). But what I find most interesting is Nvidia dropping the 4080 to a 256-bit memory bandwidth bus leaving a large 50% gap to the 384-bit bus for the 4090. One can only speculate what their plans are for a fill.

GPU​	CUDA Cores​	Memory Bus​	Memory ​
RTX 4090	16384	384-bit	24GB
RTX 4080	10240	256-bit	16GB
RTX 3090	10469	384-bit	24GB
RTX 3080	8704/8960	320-bit/384-bit	10GB/12GB
RTX 3080 Ti​	10240​	384-bit​	12GB​

 

[Western Guy]

I could write something snide about the 4090 Ti popping circuit breakers, but realistically those cards only exist to give away to YouTubers in the "have over 10 people on staff category". All of those people are running commercial power, so it really doesn't matter.

I'm a lot more upset about the idea their "4070" could be a 300W card. That's significantly higher than even a 2080Ti. I live in an area with this rare event called "Summer", which happens on a semi-regular basis. . .

A sub-200 Watt card might not even happen from Nvidia next generation, which is wild to me. If these are the results Nvidia is getting, then maybe they should finally remove the useless Ray Tracing software. Game devs don't want to support it. It's time to move on. Stop adding all that cost and power to gaming cards for tech that is really only useful to Hollywood render farms.

Commercial power? That’s a fair bit overblown.

800W is 6.67 Amps. That’s less than 50% of the power my 4 slice toaster uses. Assuming I bought the card and used it around 400 hours a year (8 hours a week) my power cost would be $22.40 a year ($0.07KW power). Plus my house has a 200 Amp service (code where I live now) which means I could run absolutely everything simultaneously and still not crack 70 Amps.

I really can’t understand when people are talking about GPUs that will cost several thousand dollars why $20 of power a year is a concern? I get your power might cost more but it’s still a tiny fraction of the overall cost. Agonizing over 100W ($2.80 a year for me) seems absurd.

 

[Bob_8_]

.,., SLI --

 

[KyaraM]

As always, I only believe any of this when Nvidia themselves release the specs. This is all speculation at this point. So many of those "leaks" feel like shots into the blue, why does anyone even still read them?

 

[Deleted member 14196]

Seems to be more for just bragging rights than a practical commercial product and to see how far they can push the limits of the technology. Kind of reminds me of a concept car, but in this case they can sell a few production units as well.

Agreed. Plus it will probably cost $10,000

 

[Flayed]

$0.07KW power

That's a good price for lecky. Round here it's 26.5p /KWh ($0.32)
For me, it's more about not wanting to dump more heat into my room. It got up to 96 F in here during the recent heatwave.

 

[King_V]

Commercial power? That’s a fair bit overblown.

800W is 6.67 Amps. That’s less than 50% of the power my 4 slice toaster uses. Assuming I bought the card and used it around 400 hours a year (8 hours a week) my power cost would be $22.40 a year ($0.07KW power). Plus my house has a 200 Amp service (code where I live now) which means I could run absolutely everything simultaneously and still not crack 70 Amps.

I really can’t understand when people are talking about GPUs that will cost several thousand dollars why $20 of power a year is a concern? I get your power might cost more but it’s still a tiny fraction of the overall cost. Agonizing over 100W ($2.80 a year for me) seems absurd.

What state is that? Checking the rates here (table defaults to Business, you have to manually switch it to Residential), the cheapest in the nation is 10.03 cents/kWh.. and that's not counting delivery.

It's also not counting the extra heat pouring into your room, and the extra air-conditioning that you need to run to compensate for it.

AND... some of us actually do care to minimize the damage we're doing, and also think efficiency is important in technological development. It's not all about the money.

 

[salgado18]

Commercial power? That’s a fair bit overblown.

800W is 6.67 Amps. That’s less than 50% of the power my 4 slice toaster uses. Assuming I bought the card and used it around 400 hours a year (8 hours a week) my power cost would be $22.40 a year ($0.07KW power). Plus my house has a 200 Amp service (code where I live now) which means I could run absolutely everything simultaneously and still not crack 70 Amps.

I really can’t understand when people are talking about GPUs that will cost several thousand dollars why $20 of power a year is a concern? I get your power might cost more but it’s still a tiny fraction of the overall cost. Agonizing over 100W ($2.80 a year for me) seems absurd.

Going beyond your own personal case:

• many people pay more than that for power;
• in some places, power cost is exponential, meaning if you cross a certain threshold, the price per kWh increases;
• that's almost the same power as a microwave heater;
• if they sell that to 1000 people, that's an extra 400kW of energy that the grid must provide; now extrapolate to even more people;
• consider the above as a tendency, and in a few years we'll need new powerplants to feed hungry GPUs just to play GTA VI;
• now the card needs water-cooling, air will hardly dissipate 800W of heat inside a closed case;
• cards will get even bigger and fatter, at least three slots;
• lastly, the heat that goes to the room must also be dissipated, and many will use air conditioning, which again uses up more energy from the grid and increase the monthly bill.

Sorry to take your post for this rant, but it had a convenient case as to why graphics cards over 400W are absurd, and should not be made the norm in any way.

 

[kal326]

Just wait 'til the RTX 5090 Ti comes out....

Come on now they will milk it for a bit more before going to that level.
So RTX 7090 Ti for that power level. Also, if you buy an EVGA FTW3 Ultra they will throw in an eLeash and Mr Fusion.

 

[_dawn_chorus_]

maybe they should finally remove the useless Ray Tracing software. Game devs don't want to support it. It's time to move on. Stop adding all that cost and power to gaming cards for tech that is really only useful to Hollywood render farms.

I would strongly disagree with this.. some games like Dying Light 2 for example look immensely better with raytracing. It looks fantastic when implemented well and while not every implementation of it looks amazing it is awesome technology.
Here is an Off/On comparison I took seconds apart:

https://imgur.com/a/dJhDK6L

View: https://imgur.com/a/dJhDK6L

 

[Phaaze88]

I would strongly disagree with this.. some games like Dying Light 2 for example look immensely better with raytracing. It looks fantastic when implemented well and while not every implementation of it looks amazing it is awesome technology.
Here is an Off/On comparison I took seconds apart:

https://imgur.com/a/dJhDK6L

View: https://imgur.com/a/dJhDK6L

They look like they were taken at different times of the day.

 

[_dawn_chorus_]

They look like they were taken at different times of the day.

Exactly except they weren't. That's how big the difference is in shaded areas. I took the first, turned on RT, restarted the game, appeared in the same place at the same time and this was the result.

You can see in the bottom right corner is the in game day time, 20 game world minutes had passed between the two. I've seen the same results just looking at my gf's screen across the room while playing coop to see the difference. She plays with RT off for higher fps.

 

[pclaughton]

That bus width and GDDR5 on the 4070 kind of surprise me.

 

[spongiemaster]

That bus width and GDDR5 on the 4070 kind of surprise me.

It's a typo. Everywhere else has the 4070 with GDDR6, with everything above it using GDDR6X. Bus bandwidth is right. Rumors are that ADA will use something equivalent to AMD's infinity cache minimizing the penalty of lower memory bandwidth.

 

[vinay2070]

So it seems like now nvidia is trying to make XX90 the new XX80 by severely undercutting the XX80 series on core count and launching them late. So when a flagship mainstream should have costed 600 to 700$ will now cost 1700$.