NVIDIA Releases Ada Lovelace-Powered RTX 40-Series GPUsSeptember 22, 2022
The highly awaited announcement is finally here! NVIDIA has added more GPUs to their arsenal—the NVIDIA RTX 4090, NVIDIA RTX 4080, and the NVIDIA RTX 6000 Ada Lovelace Generation GPUs. The new 3rdgeneration architecture on these GPUs is named after Ada Lovelace. Lovelace was a famous scientist and mathematician in the early 1800s whom historians recognize as the first computer programmer. She led the way for some remarkable technological innovations, paving the way for these new GPUs by NVIDIA.
Ada is incredibly energy efficient and leverages over twice the performance at the same power compared to Ampere. According to NVIDIA, the RTX 4090 is 2x faster on Microsoft Flight Simulator, 3x faster on Portal RTX, and 4x faster on RacerX when compared to the RTX 3090Ti, the world’s reigning GPU champion.
The new streaming multiprocessor on the RTX 4090 delivers 90 teraflops, over double compared to the previous generation. Ada’s architecture includes a major new technology called Shader Execution Reordering (SER), improving ray-tracing performance by as much as 3X. As the name implies, SER enables the GPU to dynamically reorganize shader workloads to make the most of the GPU.
Ada also introduces DLSS3, a new AI generation that generates entirely new frames, rather than just pixels. This 3rd gen deep-learning technology allows the GPU to render entire frames rather than having CPU limitations. According to NVIDIA, users can get up to 4X the performance in games compared to usual rendering methods.
Overall, the NVIDIA GeForce RTX 4090 features 16,384 CUDA Cores with up to 24GB GDDR6X memory, meanwhile the RTX 4080 supports 9,728 CUDA Cores with up to 16GB GDDR6X memory and the 12GB GDDR6X variation supports 7,680 CUDA Cores.
The initial release of NVIDIA’s new RTX 4090 Ada Lovelace GPU will be available on October 12th for $1,599, followed by the RTX 4080 in November. This version will have two variations featuring 12GB of GDDR6X memory starting at $899, while the 16GB version will set you back $1,199.
If you’re not a gamer, NVIDIA’s Ada Lovelace architecture also includes the debut of the RTX 6000 professional GPU. This new professional desktop workstation GPU is the successor to NVIDIA’s Ampere architecture. Featuring up to 48GB of memory, it delivers up to twice the performance of the previous generation RTX A6000, powered by Ampere architecture, which was released in 2020. It supports a PCIe Gen4 x16 form factor with 142 third-generation RT Cores, 568 fourth-generation Tensor Cores, and 18,176 CUDA Cores.
“Neural graphics is driving the next wave of innovation in computer graphics and will change the way content is created and experienced,” said Bob Pette, vice president of professional visualization at NVIDIA. “The NVIDIA RTX 6000 is ready to power this new era for engineers, designers and scientists to meet the need for demanding content-creation, rendering, AI and simulation workloads that are required to build worlds in the metaverse.”
The Ada-powered RTX 6000 GPU is connected to the workstation motherboard by a 16-pin PCIe power connector and has a maximum power consumption of up to 300W. With a 4nm process cooled by an active blower fan, it can be connected to 4x displays. The GPU supports NVIDIA’s 3D Vision and 3D Vision Pro via a 3-pin mini DIN and is also compatible with Quadro Sync II. The NVIDIA RTX 6000 workstation GPU is expected to be available from distribution partners and manufacturers starting in December of this year, following the release of the RTX 40-series GPUs.
If you’re interested in these, or other GPUs, we have an entire warehouse with servers, workstations, and components waiting for you! Visit the IT Creations website here, or give us a call at 1-800-237-0402.