While we are only a few days away from the launch of the GeForce RTX 4080, the first benchmarks of Nvidia’s latest GPU have just dropped. And unsurprisingly, it shows significantly better performance than the RTX 3080.
From November 16, 2022, players will be able to get their hands on the RTX 4080, one of Nvidia’s latest GPUs under the Ada Lovelace architecture. From what we know, however, it will be necessary to be particularly reactive to obtain one, because according to recent revelations, Nvidia will have planned lower stocks for the RTX 4080. The objective being, according to several leakers, to sell during the Christmas holidays the large quantity of RTX 3000 still available.
But anyway, the GeForce RTX 4080 is coming next week, and some testers have already received the beast. In fact, it was inevitable to see appear on the web the first benchmarks of the new Nvidia GPUdespite the manufacturer’s embargo.
Significantly better performance than the RTX 3080
First of all and as reported by our colleagues from the specialized site Videocardz, the GeForce RTX 4080 was tested with three different graphics APIs, namely Vullkan, OpenCL and CUDA. Here is the opportunity for us to compare its scores with previous GPUs, such as the RTX 3080 or 3090 Ti.
Before going into detail, it should be noted that the graphics card was tested with a high-end PC equipped with an ASUS ROG Crosshair X670 Extreme motherboard and a Ryzen 9 7950X CPU. We can clearly see in the OpenCL information that the bench recognized the GPU as a GeForce RTX 4080 with 9728 CUDA cores, 304 Tensor Cores, 76 RT Cores, 64 MB L2 cache and 16 GB VRAM. As for the clock frequencies, they are indeed from 2.21 to 2.63 GHz.
Also read: RTX 4080 – first benchmark gives insight into graphics card performance
Let us also recall that the RTX 4080 will show a TDP of 320 wattswhile it will be offered at the price of 1469 € in Founder Edition. But without further ado, let’s see the scores obtained:
- OpenCL: 248,932 points
- CUDA Score: 300,728 points
- Vulkan Score: 148,838 points
According to these classifications, the RTX 40980 is 45.7% faster in CUDA than the RTX 3080, 37.1% faster in OpenCL and 29.9% in Vulkan API. In comparison to the fastest Ampere GPU, namely the RTX 3090 Ti, the RTX 4080 between 5.4 and 15.5% more efficient.
But that’s not all, since we also had the right to the first scores of the GPU under 3D Mark Time Spy in 1440p Performance and in 4K Extreme. It shows results of 14,718 points and 28,599 points respectively in Time Spy and Time Spy Extreme. Fact, this is a 51-54% performance increase over the RTX 3080and by 32% against the RTX 3090 Ti in both tests.
