RTX 4080 Super: Is Nvidia finally fixing the memory problem of its graphics cards?

By
Kevin
-
0
248

The RTX 4080 is supposed to be relaunched as a Super variant. According to a leak, the specs will be slightly upgraded – especially the memory increase raises hope.

The rumors around new Nvidia graphics cards are picking up speed: After a first leak about new RTX-4000 series cards surfaced in the past few days, more concrete information is now available about at least one model.

First, the leaker, who is considered to be quite reliable, reported @hongxing2020 about three new Nvidia GPUs. This was about the RTX 4070 Super, RTX 4070 Ti Super and RTX 4080 Super. With the names, Nvidia also breaks with a previously common naming convention:

RTX 4080 Super: 20 instead of 16 GByte memory?

This original leak now seems to be confirmed, as the Taiwanese portal Benchlife reports. They refer to sources at Nvidia board partners, according to which the three RTX Super graphics cards are actually in the works.
However, concrete specifications of the upcoming GPUs could only be elicited to a limited extent, as the leaker @harukaze5719 also explains.

Only the memory size of the RTX 4080 Super is said to have already been determined. This is supposed to be 20 GBytes of GDDRX6 memory and

After all, further specifications of an RTX 4080 Super can already be derived from this capacity. Because 20 GBytes of VRAM also require a larger memory interface. Instead of 256 bits, the memory would have to be accessed via a 320 bit interface.

It also follows that a rumor from the beginning of October hit the mark: A post by the leaker MEGASizeGPU assumed that such a graphics card would have to be equipped with the AD102 chip, i.e. the same unit as in the RTX 4090.

Why does that have to be the case? Ultimately, this is due to the AD103 chip of the RTX 4080, which can’t work with 320-bit interfaces; so only an upgrade can be considered sensible here, if the Super variant is to actually sort itself between the RTX 4080 and 4090.

In this context, the portal (Videocardz) speculates that an RTX 4070 Super will take a similar upgrade path. If the new variant goes analog steps, it would mean an AD103 chip as well as 16 GBytes of memory with 256-bit connectivity.

Nothing concrete can be said about a price for the RTX 4080 Super and the other rumored models. The only clue is the MEGASizeGPU leak, which assumes a price in the range of the original RTX 4080 UVP, i.e. around 1,400 to 1,500 euro.

Editor’s Opinion: This is what RTX 4080 and 4070 should have looked like from the start

Let’s briefly recall a GPU debacle from last year: in an unprecedented move, Nvidia took the original RTX 4080 form factor off the portfolio before release. 12 GBytes of memory and a slow chip caused a lot of criticism from potential buyers, and quite rightly so.

Subsequently, the RTX 4080 was launched as the 4070 Ti, but buyers do not seem to be very happy with this variation either. Accordingly, it is first of all to be welcomed that an RTX 4080 with 20 GB and an RTX 4070 with 16 GB VRAM, respectively, are finally to be released, whatever they will be called.

The only question remains: Why not like this from the start? The development that eight to twelve GB of video memory are only conditionally future-proof depending on the resolution did not just start recently. If the two speculated Super variants had been released as regular RTX 4080 and RTX 4070 a year ago, Nvidia would have been spared a lot of trouble in my opinion.

Now it’s your turn: Is a RTX 4080 Super with 20 GBytes just the right model to convince high-end users aside from the RTX 4090? Would the RTX 4070 Super with 16 GBytes be an ideal mid-range card? What price point would the two graphics cards have to have to convince you to buy them? Leave us your opinion in the comments!