Leaked benchmarks reveal that the performance of NVIDIA’s GeForce RTX 4080 SUPER GPU closely mirrors that of the non-SUPER 4080, contrasting with the notable performance improvements seen in the RTX 4070 SUPER and RTX 4070 Ti SUPER models.
BenchLeaks discovered the leaked benchmarks on the Geekbench 5 database. It’s worth noting that the GB5 benchmark, although outdated and replaced by GB6, is used for synthetic tests and may not precisely reflect real-world gaming performance. Nevertheless, past graphics performance in this benchmark has closely aligned with gaming performance.
In the performance assessments, the NVIDIA GeForce RTX 4080 SUPER achieved 309,554 points in the CUDA test, 264,806 points in the OpenCL test, and a relatively lower 100,378 points in the Vulkan test. Given the low Vulkan score, mirroring a situation observed with the 4070 Ti SUPER, it can be disregarded. Meanwhile, the CUDA and OpenCL scores suggest that the card’s true graphics horsepower is in line with that of the RTX 4080 Non-SUPER.
Despite having a slightly higher core count, NVIDIA has acknowledged that the RTX 4080 SUPER is only expected to be 3% faster than the RTX 4080 Non-SUPER, particularly noticeable at higher 4K resolutions with minimal differences at sub-4K resolutions. With a price drop to $999 US, it becomes a compelling option against the RX 7900 XTX, but overall, gamers may not anticipate significant improvements from this final SUPER variant.
In the Geekbench 6 OpenCL test, the NVIDIA GeForce RTX 4080 SUPER demonstrates a 7% advantage over the Non-SUPER variant. It’s essential to note that these results are based on synthetic workloads, so actual gaming performance may vary slightly.
The NVIDIA GeForce RTX 4080 SUPER employs the AD103-400 GPU (PG139 SKU 355) featuring a complete set of 10,240 cores, 320 TMUs, 112 ROPs, and 64 MB of L2 cache. This amounts to a 5% increase in cores compared to the Non-SUPER RTX 4080s. Additionally, the GPU operates at slightly higher clock speeds, with a 2295 MHz base and 2550 MHz boost, in contrast to the 2205 MHz base and 2505 MHz boost of the RTX 4080 Non-SUPER variant.
The graphics card utilizes a 256-bit bus interface with 16 GB GDDR6X memory, matching the specifications of the RTX 4080 Non-SUPER but elevated to 23 Gbps. This enhancement establishes it as the world’s fastest G6X solution, delivering up to 736 GB/s of bandwidth, approximately a 3% increase over the Non-SUPER variant. The NVIDIA GeForce RTX 4080 SUPER maintains the same 320W TGP as the RTX 4080 Non-SUPER. With capabilities of 52 Shader TFLOPs, 121 RT TFLOPs, and 836 AI TOPs, the card supports the latest NV encode technologies such as AV1 and H.264.
Both Founders Edition and custom models of the NVIDIA GeForce RTX 4080 SUPER will launch on January 31, 2024, with prices starting at $999 US.
Source: BenchLeaks