• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Nvidia RTX 60 series GPU Specifications Leak

erek

Fully [H]
2FA
Joined
Dec 19, 2005
Messages
17,421

“Nvidia RTX 6090​

NVIDIA’s RTX 6090 will reportedly use NVIDIA’s GR202 silicon, which reportedly features up to 192 Stream Multiprocessor (SM) units. The GPU is also expected to feature a 512-bit memory bus, with the GPU supporting 32GB of memory.

Nvidia’s RTX 6090 is unlikely to use all of Nvidia’s GR202 silicon, much like most other RTX XX90-class GPUs. Nvidia’s RTX 5090 features 170 SM units, which means that the RTX 6090 could feature an SM count that’s up to 12.9% higher. Furthermore, Nvidia’s RTX 60 series GPUs will also reportedly feature slightly higher clock speeds than their RTX 50 series GPU models.

If these specifications are correct, Nvidia’s RTX 60 series performance gains will stem from architectural advantages and not raw increases in SM count. If Nvidia intends to deliver a 2x increase in path tracing performance, we can expect some big changes to how ray tracing is handled with RTX 60 series GPUs.”

Source: https://overclock3d.net/news/gpu-displays/alleged-nvidia-rtx-60-series-gpu-specifications-leak/
 
I can’t wait to not be able to buy these cards!
There was plenty of opportunities during the 50XX series release to purchase them at MSRP. We won't know if things will change for the better or worse upon next generations releases, but people have got to let go of this incessant negativity at next generation rumors because they somehow feel slighted by the current situation that is affecting nearly the entirety of enthusiast PC building. We sometimes forget what had it so good for so long, and like all good things; will certainly come to an end at some point.
 
- Nvidia’s RTX 6090 is expected to be a bigger GPU than its predecessor
- Nvidia’s RTX 6080 and RTX 6070 will reportedly feature memory bus sizes of 320-bit and 256-bit.

Would be surprised about those 2 things, 5090 is already 750mm
, memory bus going up instead of keeping them low and up to 3GB module would also be not what I would have expected.

6080/6070 getting big enough for 320/256 bits (maybe tsmc 3N let you do it with lovelace-blackwell smaller die but memory controller do not scale that well) would be a nice step.
 
There was plenty of opportunities during the 50XX series release to purchase them at MSRP.
Not in Europe, there wasn't. It started at 2.5x MSRP.
We won't know if things will change for the better or worse upon next generations releases, but people have got to let go of this incessant negativity at next generation rumors because they somehow feel slighted by the current situation that is affecting nearly the entirety of enthusiast PC building.
Yeah, just stop being such negative nancies, stop noticing that after nvidia climbed to the top on your backs they are now stepping on your faces and laughing at you.
We sometimes forget what had it so good for so long, and like all good things; will certainly come to an end at some point.
Like the AI bubble and having friends at high places for nvidia?
 
I'm not sure what's the point of even releasing them. With no one able to afford actual pc's anymore, they might as well just send them to data centers directly.
 
There was plenty of opportunities during the 50XX series release to purchase them at MSRP.
The 50 series didn't have every AI startup in all existence snatching up everything with silicon in it for the next 2-3 years.
. We sometimes forget what had it so good for so long, and like all good things; will certainly come to an end at some point.
So good? So being being able to buy computer components that dont get jacked up in price due to excessive demand is "so good"?
 
I'm not sure what's the point of even releasing them. With no one able to afford actual pc's anymore, they might as well just send them to data centers directly.
Oh, they are sending them to data centers. These will be the rejects only. There will be more GPUs sent out to reviewers oops, I mean shills than retail stock available at the paper launch.

Unless the bubble pops before then. But we all must be strong if that happens and not run to buy immediately. Let them cook, until they come crawling back to enthusiasts with realistic pricing.
 
I'm not sure what's the point of even releasing them. With no one able to afford actual pc's anymore, they might as well just send them to data centers directly.
There is an economic theory that the future of commerce will shift due to the emergence of AI and massive loss of jobs.

Essentially, what used to be the "middle" class (and everyone who isn't in the top 1%) will essentially become economically, commercially and industrially useless. They won't have any market value for advertising to, they won't have any useful or exploitable skills that aren't automated, and they won't have access to education to escape. Essentially the only actual "people" will be those who own assets in quantities that can produce, everyone else below them will be effectively useless. So any commercial or consumer goods would essentially be priced in a market for those top 1% of people. Much like a F2P mobile game that makes 99% of it's profit off of "whales" who represent less than 1% of the playerbase, nearly all market strategies will have to shift their strategies to pricing their products with such a margin that 1% of the population can sustain their business. Essentially an elites-only economy of people who own the means of production.

So you're not far off.

Nvidia is already prepping for this by slowly but surely pricing their top-end products further and further away from the achievable realities of most people. Imagine a market where the only products are different 6090s, all over $30,000. you'll have a 6090, a 6090 Ti, a 6095, a 6097 etc. and the 6060 will be a flake of silicon drawing 15w with 8GB of RAM that costs $799 that performs worse than a 5060 but has access to 5x Frame-Gen so Nvidia says it's faster.
 
I don’t see this happening this year, maybe mid 2027
 

“Nvidia RTX 6090​

NVIDIA’s RTX 6090 will reportedly use NVIDIA’s GR202 silicon, which reportedly features up to 192 Stream Multiprocessor (SM) units. The GPU is also expected to feature a 512-bit memory bus, with the GPU supporting 32GB of memory.

Nvidia’s RTX 6090 is unlikely to use all of Nvidia’s GR202 silicon, much like most other RTX XX90-class GPUs. Nvidia’s RTX 5090 features 170 SM units, which means that the RTX 6090 could feature an SM count that’s up to 12.9% higher. Furthermore, Nvidia’s RTX 60 series GPUs will also reportedly feature slightly higher clock speeds than their RTX 50 series GPU models.

If these specifications are correct, Nvidia’s RTX 60 series performance gains will stem from architectural advantages and not raw increases in SM count. If Nvidia intends to deliver a 2x increase in path tracing performance, we can expect some big changes to how ray tracing is handled with RTX 60 series GPUs.”

Source: https://overclock3d.net/news/gpu-displays/alleged-nvidia-rtx-60-series-gpu-specifications-leak/

So essentially like a 3070 to 4070 type of gain. Maybe with the 6090 they will push it higher, but the architectural gains in the 40 and 50 series was not that great. Unless the 60 series is going to buck the trend of the 40/50** series things don't look that great. With high RAM prices no one will be able to afford them anyways. 6090 will come out what, early 2027 at the earliest? Don't think the RAM shortage will be over, so I expect $3000+.
 
  • Like
Reactions: erek
like this

NVIDIA’s RTX 6090 will reportedly use NVIDIA’s GR202 silicon, which reportedly features up to 192 Stream Multiprocessor (SM) units. The GPU is also expected to feature a 512-bit memory bus, with the GPU supporting 32GB of memory.
$32k for 32 GB version.

Base config with 2GB so you can play at 480p like you mean it. :p
 
6070 having a 256 bit memory bus is enticing. Hopefully it'll have 16GB VRAM as I'm frequently bumping into the limits of my 5070 12GB.
 
6070 having a 256 bit memory bus is enticing. Hopefully it'll have 16GB VRAM as I'm frequently bumping into the limits of my 5070 12GB.

5070ti does so I'd assume they would try and move that down to a lower tier model but with VRAM pricing I really doubt it. Unless things change a lot by then.
 
It's just crazy how powerful their 4, 6, and 8 GB models are going to be. 5090 performance at only $1,499.
 
There was plenty of opportunities during the 50XX series release to purchase them at MSRP. We won't know if things will change for the better or worse upon next generations releases, but people have got to let go of this incessant negativity at next generation rumors because they somehow feel slighted by the current situation that is affecting nearly the entirety of enthusiast PC building. We sometimes forget what had it so good for so long, and like all good things; will certainly come to an end at some point.

My post was intended to be funny, but anyway.
 
There was plenty of opportunities during the 50XX series release to purchase them at MSRP. We won't know if things will change for the better or worse upon next generations releases, but people have got to let go of this incessant negativity at next generation rumors because they somehow feel slighted by the current situation that is affecting nearly the entirety of enthusiast PC building. We sometimes forget what had it so good for so long, and like all good things; will certainly come to an end at some point.
Other than for a few minutes during paper launch day, there was approximately an almost one-month period from mid-September to mid-October 2025 in which you could buy a non-Founder's Edition 5090 from Microcenter for roughly $2,000. A number of Microcenters tended, most days during that period, to have in stock listings for 1-3 of the lowest-tier 5090's from Gigabyte or sometimes MSI for in-store purchase only. Not Newegg. Not Best Buy. Not Amazon. Not B&H. Practically just Microcenter and some San Fransisco small chain stores.
 
- Nvidia’s RTX 6090 is expected to be a bigger GPU than its predecessor
- Nvidia’s RTX 6080 and RTX 6070 will reportedly feature memory bus sizes of 320-bit and 256-bit.

Would be surprised about those 2 things, 5090 is already 750mm
, memory bus going up instead of keeping them low and up to 3GB module would also be not what I would have expected.

6080/6070 getting big enough for 320/256 bits (maybe tsmc 3N let you do it with lovelace-blackwell smaller die but memory controller do not scale that well) would be a nice step.
the rumour is that 6080 will be a RDNA 5 AT0 killer

that is if AMD dares to release an AT0 for gamers

as of now AT0 is designed for Microsoft xcloud on a 512bit gddr7 bus to virtualize 8 series S2 (AT4) sessions on a single AT0 chip. you can think of it as approx 8x ps5 pro

if AMD releases an AT0 for gamers then it would be 384 bit gddr7 bus (6x ps5 pro) & then further cut down versions below that (320bit bus & 256bit bus)

nvidia's idea is that under no circumstances should AMD have a gaming card positioned betweened 6080 & 6090

if all else fails then the nuclear option for nvidia is to rapidly release a 2nm 6080 super refresh that can bury the AT0 (If AMD dares to release such a card)
 

“Nvidia RTX 6090​

NVIDIA’s RTX 6090 will reportedly use NVIDIA’s GR202 silicon, which reportedly features up to 192 Stream Multiprocessor (SM) units. The GPU is also expected to feature a 512-bit memory bus, with the GPU supporting 32GB of memory.

Nvidia’s RTX 6090 is unlikely to use all of Nvidia’s GR202 silicon, much like most other RTX XX90-class GPUs. Nvidia’s RTX 5090 features 170 SM units, which means that the RTX 6090 could feature an SM count that’s up to 12.9% higher. Furthermore, Nvidia’s RTX 60 series GPUs will also reportedly feature slightly higher clock speeds than their RTX 50 series GPU models.

If these specifications are correct, Nvidia’s RTX 60 series performance gains will stem from architectural advantages and not raw increases in SM count. If Nvidia intends to deliver a 2x increase in path tracing performance, we can expect some big changes to how ray tracing is handled with RTX 60 series GPUs.”

Source: https://overclock3d.net/news/gpu-displays/alleged-nvidia-rtx-60-series-gpu-specifications-leak/
complete (unauthorized) transcript of RGT's video — for people like me who hate wasting time on videos & prefer text

https://www.noobfeed.com/hardware/rdna-5-rtx-60-zen-7-intel-updates
 
is interesting, even TPU picked up this story, tech journalism is suspect,

NVIDIA Readies Rubin-based GeForce RTX 60-series with Massive RT Performance Gains

by btarunr Today, 03:02 Discuss (34 Comments)
The rumor mill has started grinding about NVIDIA's next-generation gaming GPU, and it looks like NVIDIA does not want to allow the current market environment to get in the way of implementing its product roadmap, with a roughly 2-year GeForce generation product launch cadence. The next-generation GeForce RTX 60-series will be powered by the "Rubin" graphics architecture. "Rubin" already debuted on NVIDIA's bread-winning AI GPU series, and is making its way to GeForce, reports RedGaming Tech.

The first slice of rumors about GeForce RTX 60-series predicts that NVIDIA will stick to a more conservative approach with foundry nodes, and not go with a sub-2 nm nanosheet-based node. GeForce "Rubin" will be built on some variant of the current TSMC 3 nm FinFET node. They need not be the same N3 node that's in use by Apple, Intel, and others; and NVIDIA might collaborate with TSMC on an exclusive variant just the way it created the NVIDIA 4N node, derived from TSMC N5. Chips in the series will follow the numbering scheme "GR20x," with examples being "GR202" for the biggest part powering the flagship product. The 3 nm node will allow NVIDIA to maintain GPU clock speeds ranging between high 2 GHz and low 3 GHz, which is a minor increase over the current "Blackwell."“
 
RTX 6090 for $6090 (almost the double of performance of the 5090 !!)*

Simple as !

* - In 8k with Path Tracing and DLSS 5.3 grease intense beautificator at 990w
 
I’ll follow along with my popcorn like the last few releases. Interested in seeing the power draw and how high the price will go.
 
  • Like
Reactions: Niner
like this
As pointed out above, those rumors are quite only earlys rumours, but they are directionally true If the 6090 is a bit bigger (say 770mm) and the SM count barely move despite a quite significant node upgrade to an quite updated by then TSMC 3 in 2027...

Make a bit wonder will they put quite a bit of groq technology of them (SRAM compute) in them, to run near instant inference, DLSS 6/7, ray reconstruction level of upscaling/neural rendering at near instant speed and RT cores that do more things of the raytrees/handling of rays and liberated the sm from them.

But they could get lazy and just scale Ampere again (220-230 SM).
 
nvidia's idea is that under no circumstances should AMD have a gaming card positioned betweened 6080 & 6090
If the AT0 is an 600mm+ affair, I am not sure they would have a choice, it would be quite the turn around to change the 6080 that much from the previous 2, that would either mean a 6070 big enough (well 256 bits rumours so maybe) it can replace the "5090" GB203 mobile (378mm) that was an almost exact drop-in replacement for the 4090 mobile (both 378/379mm on the 256 bits bus) or they stop that product, because I doubt a AT0 competitor would fit well in Laptops.

They have so much money they could tape out one more than usual, 770mm (6090) a "new" product in a 600mm 6080, still make the 5080 in a 400mm they call 6070, 6070 is now called 6060 and the 6050 (a 6060) become more prevalent instead of an afterthought, rise the price tag of all SKUs
 
Back
Top