Nvidia announced the arrival of 3000 series during their Ampere event held on 1st September which was addressed by the CEO, Jensen Huang. The second generation of RTX will have superior performance with improved raytracing and superior machine learning performance. After several months of rumours, Nvidia launched their new series of graphic cards with Ampere architecture.
Features of Ampere Architecture:
- Second Generation of RT cores:
New RT cores provide twice the throughput of the previous generation with concurrent ray tracing and 58 RT-TFLOPS of processing power.
- New multicores:
These multicores are the essential building block for the most efficient and fast GPU in the world. The GPU delivers twice the FP32 throughput of the previous generation.
- Third gen Tensor cores:
New dedicated tensor cores provide twice the performance making it faster and more efficient to run Artificial Intelligence-powered services like NVIDIA DLSS and provides 238 Tensor-TFLOPS of processing power.
- RTX I/O:
Enables rapid GPU-based loading and game asset decompression which accelerates input/output performance up to 100 times compared with hard drives and traditional storage APIs. In combination with Microsoft’s new DirectStorage for Windows API, RTX IO offloads dozens of CPU cores’ worth of work to the RTX GPU which improves frame rates and enables near-instantaneous game loading.
- Lightning Speed:
Nvidia worked with Micron to develop the world’s fastest discrete graphic memory for the second generation of RTX series which is known as GDDR6X that can provide 1TB/s system memory bandwidth.
Design of the Second generation of RTX Series:
Exquisite Design – A mechanical structure with a low-profile leaf spring along with a new 12-pin power connector which allows more space for components and cooling, and is compatible with 8-pin connectors in existing power supplies.
Improved Thermal Solution – The new series has 2x more cooling performance with a stunning design. Gamers and creators will be able to enjoy unparalleled performance while their GPUs simultaneously run cooler and quieter than ever before.
HDMI 2.1 – The increased bandwidth provided by HDMI 2.1 allows, a single cable connection to 8K HDR TVs for ultra-high-resolution gaming which has never been provided before
AV1 Decode – First discrete GPUs with support for the new AV1 codec which enables gamers to watch up to 8K HDR internet video using half the bandwidth.
Let’s take a look into the three graphic cards unveiled:
GeForce RTX 3070:
The 3070 is going to be sold for $499 and is equipped with 8GB of GDDR6 memory, hitting the sweet spot of performance for games running at 4K and 1440p resolutions. The RTX 3070 has superior performance than the RTX 2080 Ti at less than half the price, and on average is 0.6 times faster than the original RTX 2070.
GeForce RTX 3080:
The 3080 starts at $699 and is up to 2x faster than the RTX 2080. Featuring 10GB of the high-speed GDDR6X memory running at 19Gbps, the RTX 3080 can consistently deliver 60 frames per second (fps) for 4K gaming, making it one of the ultimate GPUs made for gaming.
GeForce RTX 3090:
The RTX 3090, priced at $1,499 is also called as the “BFGPU” — Big Ferocious GPU and it comes with a silencer — a three-slot, dual-axial, flowthrough design that is up to 10 times quieter than the TITAN RTX and keeps the GPU up to 30 degrees Centigrade cooler. The 24GB of GDDR6X memory of the GPU can tackle the toughest AI algorithms and feed massive content creation workloads. The RTX 3090 is up to 50% faster than the first generation RTX’s ultimate PC graphics card, TITAN RTX and enables gamers to experience 60 fps in 8K resolution across many performances demanding games.
Recently, NVIDIA uploaded a video of DOOM Eternal gameplay comparing the performance and graphics of RTX 2080 Ti and RTX 3080 to showcase the upgrade.
People preordering the RTX 3000 series will get a free copy of Watch Dogs: Legion and a year of subscription to NVIDIA GeForce NOW which is game streaming service.