Filter
-

GeForce RTX 3060

View:
7 Results

NVIDIA GeForce RTX 3060

LHR: Low Hash Rate as an Anti-Mining Measure

In response to the cryptocurrency boom and its impact on availability for gamers, NVIDIA has equipped its gaming graphics cards with a throttle that limits cryptocurrency mining. The performance in normal operation, especially during gaming, is not affected. An up-to-date version of the NVIDIA driver is required for operation.

Corresponding graphics card models that were subsequently equipped with a mining lock are marked with an addition in the name such as "V2" or "LHR" depending on the series and model. Some graphics card series have already been equipped with a mining lock since their introduction and therefore do not include the name addition.

The GeForce RTX 3060, introduced by NVIDIA on February 25, 2021, is a mid-range graphics card with good price-performance ratio. The GA106 graphics chip of the GeForce RTX 3060 consists of approximately 12 billion transistors, which are accommodated on only 276 mm² thanks to the 8-nm manufacturing process. Ampere includes revised computing units in the form of SMs (Streaming Multiprocessors), 3rd generation Tensor cores, and 2nd generation ray tracing cores. The throughput of these computing units could be almost doubled compared to the Turing architecture. Other features include DLSS AI acceleration, NVIDIA Reflex, NVIDIA Broadcast, and NVIDIA Omniverse Machinima. Additionally, NVIDIA has equipped the GeForce RTX 3060 with 12 GB of GDDR6 video memory.


Performance Data of the NVIDIA GeForce RTX 3060 Graphics Card (Reference Model):

    • Computing Power (FP32): approx. 12.7 TFLOPS
    • CUDA Cores (Shader): 3,584
    • RT Cores: 28
    • Tensor Cores: 112
    • Base Clock of the GPU: 1,320 MHz
    • Boost Clock of the GPU: 1,770 MHz
    • Graphics Memory: 12 GB GDDR6X
    • Memory Clock: 7,500 MHz (effective 15,000 MHz)
    • Memory Interface: 196 Bit
    • Memory Bandwidth: 360 GB/s
    • TDP: 170 Watts
    • Video Outputs: DP 1.4a, HDMI 2.1

Ray Tracing

Ray tracing enables realistic lighting through the simulation of physical properties and the behavior of light. RT is the calculation of the color of individual pixels during the path the light takes from the user's eye through a 3D landscape. The light traverses the 3D scenery and is reflected by objects or creates reflections. At the same time, shadows, transparent or semi-transparent objects are created, resulting in light refractions or distortions.

Ray tracing combines all these effects to display the final color of a pixel on the monitor. Until now, ray tracing has mainly been used in movies to create special effects. For these scenes, render farms are often used, and calculating complex 3D animations and special effects for feature films can take several weeks. Due to the complexity of ray tracing, PC applications like computer games are typically filled with 3D objects based on a raster model. In rasterization, 3D objects are modeled from a network of triangles or polygons. In the rendering pipeline, these triangles of the 3D model are individually converted into pixels on a 2D surface. The pixels can then be further processed ("shaded").


Tensor & DLSS

Ampere GPUs have fewer tensor cores overall than Turing. However, since the 3rd generation tensor cores offer four times the performance of the 2nd generation, Ampere can achieve almost twice the performance with half the tensor cores compared to Turing graphics cards. Tensor cores are specialized computing cores developed for deep learning, used for AI-based DLSS.

These matrix-based computing cores are equipped with special data paths for optimized floating-point calculation throughput and also offer increased energy efficiency. The "training" and "inference" as parts of deep learning particularly benefit from tensor cores. Training, the learning of new skills based on existing data, is intended to achieve up to 12 times higher TFLOP performance with the help of tensor cores compared to Pascal, the predecessor of the Volta GPU architecture. Inference in deep learning, the application of the skills learned during training, is intended to achieve up to 6 times higher peak TFLOP performance with Volta compared to Pascal.


The Current NVIDIA GeForce RTX 30 Graphics Cards at a Glance:

Want to know which NVIDIA graphics card is right for you? The following list includes all current models of the Ampere generation, sorted by performance, starting with the most powerful graphics card.