Alleged GeForce RTX 4060 Ti Specs Leak

Published by

Click here to post a comment for Alleged GeForce RTX 4060 Ti Specs Leak on our message forum
https://forums.guru3d.com/data/avatars/m/282/282473.jpg
Picolete:

Nvidia will compare it with the new 3060 that has lower bandwidth than the original ones
not only that, they won't tell you they're comparing with the 8g.
https://forums.guru3d.com/data/avatars/m/227/227994.jpg
Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
https://forums.guru3d.com/data/avatars/m/282/282473.jpg
TheDeeGee:

Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
yes. I'm sure lovelace with its cache will do a little better than ampere, and 4060 will probably use faster gddr6 than 3060, but nevertheless, a 128-bit bus on a 1440p card isn't optimal.
https://forums.guru3d.com/data/avatars/m/282/282473.jpg
Alessio1989:

rx7700 waiting..
just buy a 6800 now, seriously. better card, better value, comes with two games.
https://forums.guru3d.com/data/avatars/m/250/250418.jpg
220W TDP for a x060 tier of card...considered low... I want the stuff people are smoking too to cope with this stupidity.
data/avatar/default/avatar09.webp
cucaulay malkin:

just buy a 6800 now, seriously. better card, better value, comes with two games.
for sure. all depends about price and how much money you can spend.
https://forums.guru3d.com/data/avatars/m/258/258664.jpg
TheDeeGee:

Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
Not that I would know but I do remember that there was buzz about memory bus already when Nvidia went from 384 or what it was to 256bit. I believe it was Pascal or even Turing where they specifically included the note that they improved compression algorithms to stream data to and from the VRAM. That said, I have no idea if this really is a big deal, or if it was back then, in the end I just tend to look at independent benchmarks once the cards are out and judge by that.
https://forums.guru3d.com/data/avatars/m/56/56004.jpg
Undying:

Lets be realistic its not gonna be 700$ but most likely close. Something like 499$ which is a 100$ jump from 3060ti. Sounds reasonable(in Jensens head). 😀
Maybe I'm a pessimist, but I don't think we'll see a sub 400-450USD RTX 4060 (and moving forward, no RTX xx60 tier cards at 450USD or lower), the RTX 4060 Ti would prolly land around the 550USD mark, remember, nVidia is pricing everything up because Mr Leather Jackets needs more lambskin jackets to add to his wardrobe.
data/avatar/default/avatar20.webp
This is basically a 2080 with better RT hopefully. I hope it is priced sane but its not a 4K GPU.
https://forums.guru3d.com/data/avatars/m/56/56686.jpg
mohiuddin:

Hope it is priced below 700$ Finally a low-mid range card below 700$ Thanks NVIDIA
yeah it need well bellow 400$ which probably wont, and 8gbs GDDR6 @ 128bit less that has some crazy optimizing they gona have less bandwidth then 3060( 448GB/S) and probably 3050 (224GB/s), it would need GDDR6X just match or even beat those 2 card bandwidth wise seeing the 3050 is already 128bit and using both the cards are using GDDR6 and "new" 3060ti has GGDR6X with like 660 GB/s bandwidth. if those spec even remotely true, i have lost hope in nvidia and or hope of getting gpu. that ridiculously over price under whelming and using to much TDP
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
TheDeeGee:

Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
Memory bandwidth is fairly important in 2 ways: 1. If you don't have enough VRAM, where it has to empty and replenish itself as quickly as it can. 2. If you're working with high-res textures and meshes. This is pretty much the #1 performance killer for integrated GPUs. While both AMD and Intel have the potential to make a very potent iGPU, they don't bother because DRAM has never been fast enough, and it's too expensive to either add more cache or more memory channels. iGPUs tend to have plenty of compute performance left over if you just reduce texture details, and they also get a near 1:1 performance increase with faster memory. So, when you consider a dGPU has far more performance, limiting the memory bandwidth can dramatically hurt performance. The way I see it, Nvidia is probably thinking of the 4060Ti as a 1080p RT-capable GPU. In theory, if you were to lower the texture detail a notch or two, 128-bit would be perfectly adequate and the user would have little to no loss in visual fidelity. Contrary to what a lot of people here believe, you can lower texture details in modern games and have no perceptible difference at 1080p (and the only time you'd notice is if you put your face up against the wall and stare at it, in which case, you're not playing the game), because maxed-out textures in modern games are optimized for 4K displays. It's just one of those things that people like to get petty about, where just the notion of lowering a detail setting is totally unacceptable no matter how little of an impact it has. And yet, many of those people are completely fine with DLSS...
https://forums.guru3d.com/data/avatars/m/198/198862.jpg
schmidtbag:

The way I see it, Nvidia is probably thinking of the 4060Ti as a 1080p RT-capable GPU. In theory, if you were to lower the texture detail a notch or two, 128-bit would be perfectly adequate and the user would have little to no loss in visual fidelity. Contrary to what a lot of people here believe, you can lower texture details in modern games and have no perceptible difference at 1080p (and the only time you'd notice is if you put your face up against the wall and stare at it, in which case, you're not playing the game), because maxed-out textures in modern games are optimized for 4K displays. It's just one of those things that people like to get petty about, where just the notion of lowering a detail setting is totally unacceptable no matter how little of an impact it has. And yet, many of those people are completely fine with DLSS...
Nvidia fanboys https://abload.de/img/fetchimage7ai6d.jpeg
data/avatar/default/avatar02.webp
TheDeeGee:

Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
Yes. I would also think someone so knowledgeable in future vram requirements, especially when vram management is about to be fundamentally overhauled with Directstorage, would know this pretty clearly.
https://forums.guru3d.com/data/avatars/m/260/260828.jpg
TheDeeGee:

Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
The new 3060 8Gb 128bit was compared to the original 3060 12Gb 192bit, it's around 20% slower
https://forums.guru3d.com/data/avatars/m/216/216349.jpg
TheDeeGee:

Someone remind me. Why does everyone get their panties in a twist again when it comes to the memory bus? Has it been proven that it lowers FPS?
Smaller memory bus normally means a potential bottleneck and nobody likes that, specially when GPUs are so expensive...
https://forums.guru3d.com/data/avatars/m/238/238382.jpg
128bit... pass Also the shitty new power connector, double pass... Honestly older gen looks more attractive.
https://forums.guru3d.com/data/avatars/m/198/198862.jpg
H83:

Smaller memory bus normally means a potential bottleneck and nobody likes that, specially when GPUs are so expensive...
It will surely bottleneck at 1440p or higher. It has less memory bandwidth than 3060ti 😀
data/avatar/default/avatar14.webp
GPU comments are ridiculous. Everyone complaining about prices etc but they all run out and buy them anyway... maybe not immediately but certainly quick enough to always be within the last gen or two, just look at the Steam stats where the top places are dominated by **50 and **60 cards of the last couple of gens. If people were a little more content with what they had instead of constantly chasing the latest gen for the next 'big' upgrade, RTX, 4K, OLED... maybe prices would start coming down a little. It used to be the case that if you wanted a good, mid range card you bought a higher end card of the last gen and while a few still do, that principle is by far in the minority. The longer this goes on the more people will be driven out of PC gaming all together and then we'll all be on consoles.
https://forums.guru3d.com/data/avatars/m/238/238382.jpg
James Frazer:

GPU comments are ridiculous. Everyone complaining about prices etc but they all run out and buy them anyway... maybe not immediately but certainly quick enough to always be within the last gen or two, just look at the Steam stats where the top places are dominated by **50 and **60 cards of the last couple of gens. If people were a little more content with what they had instead of constantly chasing the latest gen for the next 'big' upgrade, RTX, 4K, OLED... maybe prices would start coming down a little. It used to be the case that if you wanted a good, mid range card you bought a higher end card of the last gen and while a few still do, that principle is by far in the minority. The longer this goes on the more people will be driven out of PC gaming all together and then we'll all be on consoles.
It's not a need for most, just a want.. I want a new video card but my current one works just fine, I was going to upgrade in 2019 but didn't because the price of what I was going to purchase was tripled, now it's just waiting for a good card to be released that is worth the money being asked of it, I was hoping the 4060 would fill that role. Funnily enough I actually did get a Switch.