NVIDIA’s most recent addition to their Ada Lovelace portfolio is the GeForce RTX 4080. The RTX 4080 has 16GB of GDDR6X memory and the AD103 GPU. Although this GPU won’t be on sale for another 10 days or so, benchmarks showing the performance you may anticipate have already leaked. The RTX 4080 has been evaluated on Geekbench using the CUDA, VULKAN, and OpenCL APIs.
The NVIDIA GeForce RTX 4080 is scheduled to be on sale next week, but Geekbench 5 tests for CUDA, OpenCL, and Vulkan seem to have already been performed on the graphics card. The graphics card was evaluated on a test system powered by an AMD Ryzen 9 7950X processor and equipped with 32 GB of DDR5 RAM and an ASUS ROG Crosshair X670E Extreme motherboard.
The NVIDIA GeForce RTX 4080 graphics card achieved a performance rating of 300728 points in the CUDA benchmark test, 248932 points in the OpenCL benchmark test, and 148838 points in the VULKAN benchmark test. When compared to the RTX 3080, this is a significant improvement of 45%. It continues to maintain its lead across the board due to the fact that the RTX 4090 is its sole genuine rival in this context.
The RTX 4080 achieves a score of 289014 points on VULKAN. It is estimated that the RX 7900 XTX on display here is 1.7 times quicker than the 6900XT. Although it may not be the ultimate outcome, it can help us get a general notion of where RDNA3 is at. In spite of costing more, the RTX 4080 here falls short of the 7900XTX.
The RTX 4080 barely outperforms the RTX 3090 in OpenCL, scoring 148838 points. Intriguingly, AMD’s 7900XTX does better in this test than even the RTX 4090. In comparison to the RTX 3080, the RTX 4080 boasts a 7% generation-on-generation boost. Even still, the whole narrative will be revealed in the final evaluations, which is sort of upsetting.
In comparison to the previous version, which provided 80 SMs or 10,240 cores, the NVIDIA GeForce RTX 4080 16 GB graphics card is anticipated to use a scaled-down AD103-300 GPU architecture with 9,728 cores or 76 SMs activated of the total 84 units. Due to its condensed architecture, the RTX 4080 could only have 48 MB of L2 cache and fewer ROPs than the complete GPU, which has 64 MB of L2 cache and up to 224 ROPs. In terms of memory specifications, the GeForce RTX 4080 is predicted to use 16 GB of GDDR6X memory, clocked at 22.5 Gbps across a 256-bit bus interface. This will provide bandwidth of up to 720 GB/s.
According to the leaked performance data, AMD’s $999 US Radeon RX 7900 XTX or even the $899 RX 7900 XT graphics cards will have a difficult time competing against NVIDIA’s $1199 US GeForce RTX 4080 graphics cards. It won’t be until the middle of December that we will know for sure how the performance compares since the graphics card comes nearly a month before AMD’s introduction.