News Nvidia Reveals RTX 4060 Ti, 4060 with Prices Starting at $299

[qwertymac93]

Higher speed VRAM makes it more difficult to do stuff like clamshell memory designs. The fact is that since the GTX 600-series, there have been very few GPUs with memory on both sides of the PCB. Off hand, I couldn't even name the last time we had an AMD consumer graphics card that had VRAM on both sides.

We'll just have to wait and see if AMD does any "clamshell" offerings I guess. Because we don't know for sure exactly how much it increases the BOM. But like I said, a $50 increase in BOM translates to $100 retail in general. Was it a $50 increase, or only a $30 increase? I couldn't say. But there's also market differentiation and economics, so if Nvidia thinks it can charge $100 extra and people will pay it, the smart business decision is to charge $100 extra.

Keep in mind that the double VRAM will probably be most beneficial for AI workloads. There are things that you simply can't do in AI with less VRAM. Several tools (Whisper and Text Generation) end up VRAM and CPU constrained rather than GPU constrained. There will absolutely be some AI people that want more than 8GB but don't want to pay the $1200 for a 16GB 4080 or $1600 for a 24GB 4090 — especially if they can make do with a $500 4060 Ti 16GB.

You're absolutely right. Higher memory speeds place greater requirements on memory controller designs and PCB layouts to get things nice and happy with doubled up RAM modules. This can certainly cause a rise in costs for those parts. It will certainly be interesting to see if the PCBs happen to be exactly the same between the 8GB and 16GB models.

There are many, very smart people working at Nvidia. I've no doubt a lot of thought went into the pricing, segmentation, and positioning of each product! For example, having  two 128-bit dies releasing side by side with such small differences in overall size shows that costs have shifted greatly. No way could Nvidia or AMD justify all the expense and man-hours to tape out another die on the same architecture just to save ~30mm² back in the 28nm days. The decision to go with this rather than a 160-bit AD106 (or 96-bit AD107) shows that TSMC's 4nm process must be extraordinarily expensive.

Edit:
The greater VRAM model will definitely be most useful for AI and rendering use. The recent bout of games falling on their face with 8/10GB cards is largely an issue of developers just not doing a good enough job optimizing for those capacities. Still, having at least 12GB of VRAM affords the gamer with the chance of side-stepping the issue since that's about the same amount of usable VRAM available in the PS5(the primary development target for modern games).

 

[Samduhman]

Price is still to high, especially for the lower end of the series. Pre 3000 series, $299 was my price point for best bang for the buck mid-level card that I could play games at high to ultra visuals for about 3 years. Even if the 4060 can do this the current prices have most likely turned off a lot of casual PC gamers from upgrading. I'm a PC gaming addict and I would not have upgraded for a couple more years if my 3070 didn't die on me.

 

[MrN1ce9uy]

I simply cannot find a single Nvidia GPU that I'm willing to spend money on. Their flagship is the only GPU that has the vram I would be willing to pay for, yet it is priced double what I'm willing to pay. Therefore, Nvidia does not have a single "good" GPU IMO.

And no, I haven't been on the "more VRAM is always better" bandwagon until RTX 30 series was selling the RTX 3060 with more vram than the RTX 3080. Coming from a GTX 1080 Ti where I saw games using over 10GB of vram in 2019, do you really think I want a 10GB GPU in 2020 or even 2022 in the form a an RTX 3080 10GB? No effing thanks. Oh, you made a 12GB version of the 3080? Unfortunately for gamers, cryptomining demand increased the market price and now we are stuck at the market price because NVIDIA!!!! chose to slap that on as their MSRP... There's no one to blame except them for that. I don't know why anyone would stick up for them saying they put a lot of thought into it. They simply enjoyed the cryptomining boom and want to stay there when in reality the cryptomining boom is simply no more. Now they are taking pay cuts and getting trashed because they aren't selling as many GPUs at the absurd prices they chose for their MSRP. It doesn't take a friggin' genius to figure it out. You doubled the price on just about every GPU you sell, and now you're not selling as many GPUs. I guess you don't care. Now, you have another whole generation (except flagship GPU that is double the price it should be) that has inadequate specs for the long-term usage MOST gamers are looking for. Once again, you don't offer a single "good" GPU.

A "good" GPU has the following attributes:

1. Performance
2. Pricing
3. Future-proofing

There is not a single card from Nvidia that hits all three of these attributes.

 

[BoggledBeagle]

I simply cannot find a single Nvidia GPU that I'm willing to spend money on. Their flagship is the only GPU that has the vram I would be willing to pay for, yet it is priced double what I'm willing to pay. Therefore, Nvidia does not have a single "good" GPU IMO.

.....

A "good" GPU has the following attributes:

1. Performance
2. Pricing
3. Future-proofing

There is not a single card from Nvidia that hits all three of these attributes.

Who cares about what you are willing to buy, it is a stricly personal matter.

4090 is objectively the most performant card on the market, with excellent performances per dollar and per watt, and 24 GB of VRAM is enough for a few years to come as well.

4070 has also excellent performances per dollar and per watt, unfortunately with its 12 GB it has questionable future-proofing.

 

[MrN1ce9uy]

Who cares about what you are willing to buy, it is a stricly personal matter.

4090 is objectively the most performant card on the market, with excellent performances per dollar and per watt, and 24 GB of VRAM is enough for a few years to come as well.

4070 has also excellent performances per dollar and per watt, unfortunately with its 12 GB it has questionable future-proofing.

The RTX 4090 is the MOST expensive GPU per frame per dollar. It costs over double per frame (in general) than the RX 6950 XT at $599. Maybe if you cherry pick games based on fringe features and settings you could see it from a different light. All I had to do is find the first video I came to with the first game I clicked on for performance to find the numbers I was looking for. Meaning, it's very obvious to see this. By and large the raw performance shows Nvidia GPUs are a terrible value, with the best value GPU they offer (the RTX 4070) being halted in production.

GPU	Brand	Cost	VRAM	Avg FPS	Cost per frame
RTX 4090	Nvidia	$1,599​	24GB	183​	$8.74​
RTX 4080	Nvidia	$1,199​	16GB	161​	$7.45​
RTX 4070 Ti	Nvidia	$799​	12GB	141​	$5.67​
RTX 4070	Nvidia	$599​	12GB	119​	$5.03​
RX 7900 XTX	AMD	$949​	24GB	179​	$5.30​
RX 7900 XT	AMD	$779​	20GB	162​	$4.81​
RX 6950 XT	AMD	$599​	16GB	144​	$4.16​

Courtesy of Hardware Unboxed -
View: https://www.youtube.com/watch?v=DNX6fSeYYT8&t=475s

 

[BoggledBeagle]

The RTX 4090 is the MOST expensive GPU per frame per dollar IN ONE GAME, ONE SETTING, ONE RESOLUTION.

I added something important you ommited. What about some other game? In 4K? Ultra settings? Ray tracing???

And more - what about using that extra features you payed for, like upscaling and frame generation?

 

[MrN1ce9uy]

I added something important you ommited. What about some other game? In 4K? Ultra settings? Ray tracing???

And more - what about using that extra features you payed for, like upscaling and frame generation?

I'll give them the RTX 4070 is the best GPU Nvidia offers at this point in time (given my criteria for best GPU 1. Performance 2. Price 3. Future Proofing; unfortunately, it lacks in the future proofing department and doesn't beet the 6950XT in price or performance. Nvidia simply does not win any segment of the gaming GPU market IMO, which is only due to AMD adjusting pricing accordingly. If only Nvidia would do the same. Even then, the future proofing on the 4070 Ti and lower is hot garbage (minus the 4060 Ti 16GB.)

 

[MrN1ce9uy]

Funny how this thread went silent since the RTX 4060 Ti reviews released.

 

[InvalidError]

Funny how this thread went silent since the RTX 4060 Ti reviews released.

Not much left to discuss once the anticipated disappointment is confirmed.