Geforce RTX 4070: Confusion about specifications

Geforce RTX 4000: Nvidia clarifies important questions about power supplies, wattage and Co.

from Valentin Sattler
There is currently a lot of confusion about the technical data of the Geforce RTX 4070: The leaker @kopite7kimi has withdrawn an earlier forecast. In general, Nvidia’s unusual naming could also lead to problems in the future.

So far, Nvidia has only offered two graphics cards from the RTX 4000 generation: the Geforce RTX 4090 and the somewhat slower Geforce RTX 4080 with 16 GB of memory. The RTX 4080 with 12 GB of memory and a slower GPU that was announced at the original presentation was withdrawn by Nvidia because of the misleading naming. According to the current status of the rumors, it will soon be launched as the Geforce RTX 4070 Ti.

Geforce RTX 4070 – different than expected

However, the lower product portfolio is apparently shifting due to the regrouping from the RTX 4080 12 GB to the RTX 4070 Ti. Specifically, according to the leaker @kopite7kimi active on Twitter, the Geforce RTX 4070 should no longer appear in the originally predicted form based on the GPU AD104-275. This should according to the same leaker have 7,168 shaders and 10 GB of memory on a 160-bit interface.

Recommended Editorial ContentThis is where you will find external content from [PLATTFORM]. To protect your personal data, external integrations are only displayed if you confirm this by clicking on “Load all external content”:I consent to external content being displayed to me. Personal data is thus transmitted to third-party platforms. Read more about our privacy policy.

According to the 3D Center website, an expansion stage with other technical data is now circulating online, which provides for the Geforce RTX 4070 12 GB with a 192-bit interface. However, the number of shaders is unclear: The Geforce RTX 4070 Ti should appear with 7,680 shaders, and Nvidia will hardly come much closer to this value. 3D Center expects even fewer shaders for the Geforce RTX 4070 instead, specifically around 6,400.

See also  Nintendo Direct is reportedly coming this month, with Zelda: Wind Waker and Twilight Princess

This would mean a 20 percent difference in computing power to the Ti model, with ideal scaling and an identical clock. For comparison: The RTX 3070 and the RTX 3070 Ti are only ten percent apart in the PCGH performance index. The gap would be unusually large. On the other hand, it is also unusual that Nvidia (necessarily) starts so early with a Ti model. This should lead to exciting solutions, especially in the future, when Nvidia will insert the expected, slightly faster refresh graphics cards.

Just as exciting: Geforce RTX 4070 Ti: GPU photographed, probably half the size of the AD102 of the RTX 4090

So, if you trust @kopite7kimi’s statement, little is clear at the moment – except that there are changes compared to the original GPU data. As usual, only Nvidia can provide clarity here. It is quite conceivable that the company will start with the RTX 4070 disruptive fire in response to the upcoming launch of the RX 7000 graphics cards. This will hopefully reveal tangible details about the company’s next graphics cards.

Source: via 3D center