Nvidia has updated the spec sheet for their upcoming GeForce RTX 4070 SUPER GPU, fixing a previously reported error in their reviewer's guide and marketing materials. The fix came as Nvidia prepares to launch the RTX 40xx SUPER series at CES 2024. Initially, the specifications for the RTX 4070 SUPER's AD104 GPU suggested a 20% increase in the number of cores over the standard RTX 4070. Confusion arose due to of a discrepancy in the reported L2 cache size.
The first materials listed it at 36MB, which conflicts with previous leaks from 2023 that indicated a 48MB cache, a 33% increase from the RTX 4070. Nvidia has now made it clear that the GeForce RTX 4070 Super does indeed have a 48MB L2 cache. The AD104 GPU used in the graphics card supports this cache size, which significantly improves its performance.
The GeForce RTX 4070 Super uses the AD104 graphics processor, which consists of 7,168 CUDA Cores and a 192-bit memory bus. While it doesn't fully utilize the AD104 GPU like the RTX 4070 Ti, it does take advantage of the full 48MB L2 cache, which improves memory bandwidth. The original misunderstanding about the cache size remains unclear. Nvidia's official product description now correctly states 48 MB of L2 cache and a memory subsystem with a peak bandwidth of 504 GB/s.
In terms of performance, the GeForce RTX 4070 Super is comparable to the RTX 4070 Ti, offering 35.5 FP32 TFLOPS against the Ti's 40.09 FP32 TFLOPS. With a similar memory configuration, the RTX 4070 Super's performance approaches the Ti model, with the potential for further improvement through overclocking. In addition to the technical correction, Nvidia has also updated the official product sheet, which now accurately states the correct L2 cache size of 48 MB and total memory bandwidth.
These are important details for potential buyers who want to compare the performance between different GPU models and make the best decision based on their specific needs. Nvidia's correction in the RTX 4070 SUPER specifications underscores the company's commitment to providing accurate and reliable information to consumers. Although the initial error in the specifications may have caused some confusion, Nvidia has worked quickly to rectify the situation and ensure that their customers have access to the most accurate information about their products.
Through this patch, Nvidia has also emphasized the importance of L2 cache size to a GPU's performance. By increasing the L2 cache size from 36MB to 48MB, Nvidia has significantly improved the GeForce RTX 4070 Super's performance, making it a more attractive choice for those looking for a high-performance GPU.