Introduction

Today, we have with us the EVGA GeForce RTX 3060 Ti FTW3 Ultra, the company’s most premium custom-design RTX 3060 Ti Ampere graphics card. EVGA combines NVIDIA’s new performance-segment GPU with its latest-generation iCX3 Interactive cooling technology. The FTW3 Ultra also gets EVGA’s highest state of tuning for the RTX 3060 Ti, its most aesthetically-appealing cooling solution, and an innovative cooler-PCB combination. EVGA’s board design is surplus to the thermal requirements of the RTX 3060 Ti, letting its designers create an aggressive fan-curve that prioritizes low noise, while a proprietary network of sensors on the card lets it accurately respond to thermal loads for effective cooling.

The new GeForce RTX 3060 Ti by NVIDIA is designed to offer extreme quality gaming at 1440p resolution with RTX raytracing enabled, as well as high refresh-rate competitive e-sports gaming at 1440p and 1080p. NVIDIA’s design goal with the new Ampere family of GPUs has been to vastly improve the performance of real-time raytracing, bringing it to more segments. NVIDIA claims that the RTX 3060 Ti beats the previous-generation RTX 2080 Super, a $700 high-end graphics card, but at a much lower starting price of just $400. The GeForce Ampere architecture heralds NVIDIA’s 2nd generation RTX real-time raytracing technology that combines traditional raster 3D with raytraced elements, such as lighting, reflections, shadows, and global illumination.

The GeForce RTX 3060 Ti is based on the same 8 nm “GA104” silicon as the RTX 3070 launched in October. It features 38 out of 48 Ampere streaming multiprocessors physically present on the silicon, yielding 4,864 CUDA cores, 152 Tensor cores, 38 RT cores, 152 TMUs, and 80 ROPs. Its memory subsystem is unchanged from the RTX 3070 (or even the RTX 2060 Super, for that matter). It features 8 GB of GDDR6 memory running at 14 Gbps, over a 256-bit wide memory interface, achieving a memory bandwidth of 448 GB/s. The 2nd generation RTX technology combines new Ampere CUDA cores that offer concurrent FP32+INT32 math capability with 2nd generation RT cores, which offer higher performance, and the new 3rd generation Tensor core that leverages the sparsity phenomenon in deep-learning neural nets to improve AI inference performance significantly. NVIDIA leverages AI for its RTX denoiser and DLSS performance enhancement.

As mentioned earlier, EVGA GeForce RTX 3060 Ti FTW3 Ultra comes with the company’s highest factory-overclocked speeds for the silicon, with the GPU Boost frequency dialed up to 1800 MHz compared to the 1665 MHz reference. This is backed by a PCB featuring a dual- 8-pin PCIe power input configuration, higher power limits, and an innovative air-through PCB construction. Unlike many other custom-design RTX 30-series cards, the EVGA card sees the PCB length roughly match the card’s overall length, meaning you get power connectors right at the edge of the card, where you’d want them. However, large holes have been punched into the PCB and the backplate to let airflow from the third fan of the cooler flow through. Its dual-slot iCX3 cooler features an aluminium fin-stack heatsink with a number of copper heat pipes, ventilated by a trio of fans.