One of its users managed to benchmark Nvidia’s now-cancelled RTX 4080 12GB graphics cards, according to a post in the Chiphell forums. With 3DMak Fire Strike and TimeSpy results that barely matched Nvidia’s Ampere-based RTX 3080 12GB, performance was regrettably quite subpar for an RTX 4080-class GPU. Given that the RTX 4080 12GB was intended to replace the RTX 3080 12GB, its performance is disappointing.
The precise RTX 4080 12GB model that was tested was never made known by the original source. We still have a clear understanding of how these cards operate because the poster provided us with the whole GPU monitoring data, including core clocks, thermals, and power.
The RTX 4080 12GB, along with the 4080 16B and 4090, was Nvidia’s entry-level Ada Lovelace GPU when it was first unveiled at GTC. With an AD103 device, 9728 CUDA cores, a 256-bit wide bus, and 22.6Gbps of GDDR6X memory, the 4080 12GB was capable of 717GB/s of bandwidth.
However, due to the RTX 4080 12GB model’s extremely unclear naming compared to the RTX 4080 16GB, Nvidia discontinued it after receiving harsh criticism from the community. Despite the labels just implying a capacity difference between the cards, the 4080 12GB has drastically different core specs than the 16GB version.
Nvidia’s “unlaunch” occurred after RTX 4080 12GB cards began leaving the assembly lines, which is how the Chiphell user was able to obtain a card.
This undisclosed RTX 4080 12GB model was clocked at a memory speed of 1,313 MHz and a real-world boost rate of 2,820 MHz, according to the GPU-Z image from the Chiphell forum post. Thermals peaked at a temperature of 72.3C, and hot spots reached 91.2C. The maximum fan speed was 43%, however the RPM is unknown. At its highest, power usage can be seen to reach 261W.
However, the 4080 12GB’s 3DMark results speak for themselves. The RTX 4080 12GB tested only managed to achieve GPU scores of 13,472 in 3DMark Fire Strike Ultra and 10,794 in TimeSpy Extreme, according to the Chiphell report.