Benchmarking of Nvidia RTX 6000 Ada GPU shows 72% improvement over A6000 Ampere in 3DMark tests

A Reddit user by the name of Healthy-Blood-54 has shared photos of a Nvidia RTX 6000 Ada graphics card that has been taken apart. The user also included performance results from the 3DMark Time Spy benchmark.

The graphics cards were not purchased for gaming purposes but instead were bought to be used in display boxes for interactive games in visitor areas. The cards were bought at a higher price of $7,400 each, compared to the retail cost of $6,800.

Nvidia RTX 6000 Ada GPU 3DMark Time Spy revealed

In the graphics section of 3DMark Time Spy, the graphics card received 30518 points, a 72% improvement over its predecessor, the RTX A6000 Ampere, and a 2x improvement over the Quadro RTX 6000, which utilized the Turing architecture.

The card is definitely slower than the RTX 4090, but it also comes with lower clock speeds and doesn't do well in gaming benchmarks.

During full load, the user reports that the card reached 85C and that the fans settled at around 70%. The A6000, which also has a cooler that looks like a blower, was no louder than the card. Due to its workstation-hardened design, the card can operate continuously at 95°C for at least ten years without experiencing any issues.

It's interesting to note that the card had a much higher score before the Windows update, which should make it faster by more than two times.

Nvidia RTX 6000 Ada GPU specifications

The Nvidia RTX 6000's specifications are also superior to those of the GeForce RTX 4090. It has only two disabled SMs at 142, providing up to 18,176 CUDA cores. That is 11 percent more SMs and cores than the RTX 4090 GPU. As reported, RTX 6000 operates at 2.50GHz. 

The RTX 4090's 24 GB of VRAM is doubled by the card's 48 GB (EEC) boost clock, which is the same as the RTX 4090's. Although the gaming graphics card uses 21 Gbps dies, the card uses slower 20 Gbps GDDR6X memory.