Intel says Arc A770 is faster at processing ray tracing than RTX 3060

Despite an unwelcome start, Intel wants to show that this time it has a lot to say in the graphics processing sector. The CPU giant has published an official comparison in which you can see that, depending on its version, the graphics Arc A770 is faster at processing ray tracing than the RTX 3060 from NVIDIA.

As we said yesterday, Intel targeting NVIDIA instead of Radeon makes all the sense in the world, as AMD’s support for ray tracing still lags far behind the green giant’s offerings. This makes the rival to beat, of course, NVIDIA, which at least until the appearance of Intel Arc graphics has shown to be at least two generations ahead of the competition.

Intel has decided to employ dedicated hardware to process ray tracing via the DirectX Raytracing 1.0, DirectX Raytracing 1.1, and Vulkan RayTracing APIs. At the hardware architecture level, each Xe core has a Thread Classification Unit (TSU) and a Ray Tracing Unit (RTU). The technology used by the company allows it to process ray tracing asynchronously, so in that sense it is closer to NVIDIA than to AMD, which processes ray tracing synchronously (it depends on the shaders). With this on the table, Intel has technological reasons to compare itself to NVIDIA.

As far as can be seen, the graph Arc A770 is altogether 14% faster than the RTX 3060 Based on Intel’s own test results when running games at 1080p. However, one can see at least four titles where the NVIDIA GPU winswith special mention Hitman 3, where the RTX 3060 has scored 68 frames per second on average compared to 57 of the Arc A770, and Battlefiled V, which has given results of 81 and 67 respectively.

The RTX 3060 has a total of 28 cores for ray tracing, while the Intel Arc A770 has not yet mentioned all its official specifications, but it is estimated that it uses a full ACM-G10 graphics processor with a total of 32. Xe cores, so taking the data exposed in this entry would have the same amount of RTU.

Another data shown by Intel is the enhancement provided by XeSSits supersampling technology aimed at competing with NVIDIA’s DLSS, when mitigating the performance impact of ray tracing on a resolution of 1440p. At general levels, the improvement is appreciable, doubling the performance with Ghostwire Tokyo and Hitman 3.

Intel, with these internal tests that it has published, intends to show muscle and that it is capable of contributing things in a sector that eluded it a few decades ago. With NVIDIA enjoying a clear dominant position in the industry and Radeon slipping away from technologies, the presence of a fit Intel is more than necessary to generate competition and benefit consumers.