A recent report has shed light on the preliminary specifications of Nvidia’s highly anticipated GeForce RTX 4070 Super graphics card. Originally, it was believed that the GPU would feature 48MB of L2 cache. However, Nvidia quickly rectified a typo in its launch materials, confirming that the actual L2 cache size is 36MB. Despite this minor correction, the GeForce RTX 4070 Super’s advanced features make it an impressive contender in the market.
The GeForce RTX 4070 Super is armed with the AD104 graphics processor, boasting 7,168 CUDA cores and a 192-bit memory interface. While it falls short of the 7,680 CUDA cores and 192-bit memory interface found in the full AD104 GPU utilized by the GeForce RTX 4070 Ti, it compensates with a substantial 48MB L2 cache. This not only enhances its memory bandwidth capabilities but also showcases Nvidia’s commitment to optimizing performance.
The compute performance of the GeForce RTX 4070 Super is already on par with its counterpart, the GeForce RTX 4070 Ti, offering 35.5 FP32 TFLOPS compared to 40.09 FP32 TFLOPS. With the same memory subsystem, the GeForce RTX 4070 Super has the potential to delivering similar levels of performance. Furthermore, the overclocking potential of today’s AD104 GPUs is expected to surpass those produced in late 2022 or early 2023, due to constant refinement of process technologies by Nvidia and its manufacturing partner, TSMC.
One of the most exciting prospects lies in the behavior of custom GeForce RTX 4070 Super graphics cards from Nvidia’s AIB models. These boards are expected to feature enhanced TGP (Total Graphics Power) and likely possess GPUs with higher overclocking potential. This suggests that enthusiasts can further bridge the performance gap between the GeForce RTX 4070 Super and the GeForce RTX 4070 Ti through overclocking.
As Nvidia continues to innovate and push the boundaries of graphics technology, the GeForce RTX 4070 Super stands as a testament to their commitment to delivering exceptional performance and unlocking the full potential of enhanced graphics cards.
The source of the article is from the blog elblog.pl