Last week, NVIDIA announced the release of the GeForce RTX 4070 SUPER graphics card. The data sheet stated a 20% increase in core count compared to the RTX 4070, although the L2 cache was still stated as 36 MB. NVIDIA has now confirmed that the 36 MB figure was an error and has updated the product pages with the correct cache size of 48 MB, which is identical to that of the RTX 4070 Ti GPU.
A larger L2 cache is of great importance for a GPU like the GeForce RTX 4070 SUPER. This graphics card is designed for high-performance gaming PCs that offer high refresh rates at 1080p and 1440p. When using high-resolution textures and generally higher resolutions, bottlenecks may occur in the memory subsystem. Although the RTX 4070 SUPER is equipped with the latest GDDR6X memory, there may still be situations in which the memory is overloaded and leads to a loss of performance when gaming. A large L2 cache in close proximity to the GPU enables fast data transfer across all GPCs.
The decision to have 48 MB of L2 cache on the NVIDIA GeForce RTX 4070 SUPER GPU would certainly be very advantageous and would represent a significant improvement over the non-SUPER variant.
NVIDIA GeForce RTX 4070 SUPER GPU specifications: Upgraded AD104 graphics processor with 20% more cores
The NVIDIA GeForce RTX 4070 SUPER will be available with the AD104-350 or AD103-175 GPU (PG141 SKU 335), with both variants featuring 7,168 cores. The GPU is expected to have 36 MB L2 cache. The card will also be equipped with 12 GB of GDDR6X memory and the same 192-bit bus interface. The RTX 4070 SUPER will have a higher TBP (Total Board Power) of 220W compared to the existing 4070.
The card offers 36 shader TFLOPs, 82 RT-TFLOPs and 568 AI-TOPs and supports all the latest NV encoding technologies such as AV1 and H.264. Overall, this places the RTX 4070 at 95% of the performance of the 4070 Ti.
NVIDIA GeForce RTX 4070 SUPER GPU performance: Faster than an RTX 3090 and only 5% slower on average than a 4070 Ti.
The NVIDIA GeForce RTX 4070 SUPER is expected to be around 15% faster than the RTX 4070 Non-SUPER in terms of gaming performance, making it the biggest improvement in the SUPER range. According to NVIDIA, the card will also be faster than an RTX 3090 without frame generation, although performance may vary depending on the game. In addition, NVIDIA confirms that the card is only 5% slower than an RTX 4070 Ti that sold for 750EURO. This has also been replaced by the RTX 4070 Ti SUPER.
The GeForce RTX 4070 SUPER and RTX 40 SUPER series are strongly focused on artificial intelligence. With DLSS 3.5, they offer up to 2x faster performance, 70% faster image generation and 50% faster video generation through ML optimizations via the Tensor RT and Tensor RT LLM stack. These features are available on Windows and are continuously updated.
NVIDIA GeForce RTX 4070 SUPER GPU energy efficiency: Absolutely groundbreaking performance per watt.
Another important aspect of the NVIDIA GeForce RT 40 SUPER GPUs, such as the RTX 4070 SUPER, is their high power efficiency. Although the GPU is rated at 220W, the actual power consumption during gaming is only around 200W on average. This is less than half the TGP of the RTX 4090 and significantly lower than the consumption of the Radeon RX 7900 series. Even the 7800 XT has a much higher rated power consumption of 263W TBP. The video playback power is also 16W by default, while the idle power is 11W. These are aspects that the AMD RDNA 3 GPUs struggled to fix during their launch.
Starting January 17, the NVIDIA GeForce RTX 4070 SUPER GPU will be available in Founders Edition and various custom variants. The starting price is 659 euros RRP.
Source: NVIDIA
16 Antworten
Kommentar
Lade neue Kommentare
Veteran
Urgestein
Neuling
Urgestein
Urgestein
1
Urgestein
Veteran
Urgestein
Urgestein
Veteran
Veteran
Veteran
Urgestein
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →