Intel is promising a big improvement in the performance of the integrated GPUs on its chips.

As with current Gen9 GPUs, Gen11 is arranged in blocks that combine execution units (EUs: execution units) with dedicated 3D hardware, such as texture samplers.

Meanwhile, Gen11 has up to 16 EUs per block and the configuration will have 4 blocks (instead of 3 like Gen9), for a total of 64 EUs instead of 24 EUs as before.

This new GPU will use tile-based rendering, where the image will be divided into tiles so that they can all be rendered separately.

This chip will contain common multimedia blocks with encoders and decoders dedicated to motion video.

Details about Intel's new generation integrated GPU, the power to break the TeraFLOPS limit

This new GPU will support HDR displays and also Adaptive Sync technology.

What results will this increase in GPU power bring?

Details about Intel's new generation integrated GPU, the power to break the TeraFLOPS limit

This new integrated GPU will ship in 10nm processors launching in 2019, so it will most likely be incorporated in the new Sunny Cove CPU cores.

Intel also reiterated its plans to create a discrete GPU with a new architecture (called “Xe”) with a variety of configurations in terms of performance and power consumption (from integrated GPUs to components).

Refer to ArsTechnica