One of the things that we didn’t need to know as information before the NVIDIA RTX 3060 Ti launched is your IT performance, that is to say, running algorithms that are not part of the graphics pipeline. Either in a completely computation-based environment, such as in asynchronous computation, run Compute Shader programs simultaneously with rendering in games.
NVIDIA RTX 3060 Ti performance under Geekbench
As you can see from the graph, the computing performance of the new NVIDIA graphics card is very close to that of AMD’s RX 6800.
How can this happen? Let’s not forget that each SM, equivalent to AMD Compute Units, of all GeForce Ampere has 128 FP32 ALU, while The AMD architecture of Compute Unit has “only” 64 ALUs. So you are comparing a GPU with 4864 ALU with 32-bit floating point precision, RTX 3060 Ti, with one that has 4096 ALU of the same precision.
The advantage of the RX 6000 lies in the non-duplicated units such as fixed functional units, which also have the advantage of a higher clock speed, while the advantage of the RTX 3060 Ti and its sisters is all that has to do with computing through the GPU.
What does this mean in terms of performance?
While waiting for the NVIDIA RTX 3060 Ti tests, stay tuned to the one we’re posting this afternoon on HardzoneWe have to keep in mind that NVIDIA has taken a different path with its architecture, instead of preparing it more for existing games where its GPUs are already performing well, optimized it for future games
A Compute Shader is the concept of using the power of the GPU to solve problems outside of rendering, But paradoxically they are used to achieve image post-processing effects in games Yes are important for the implementation of Ray Tracing, since types of shaders used inside they are nothing more than Calculate shaders.
This is why computational performance is very important, because more and more games use less shaders depending on fixed function units. Obviously, not all games use GPU resources in the same way, but games that use Ray Tracing, which is one of NVIDIA’s spearheads, use and benefit from GPU computing.