Intel spoke about the monstrous power of the flagship Xe-HP graphics on four crystals

Intel spoke about the monstrous power of the flagship Xe-HP graphics on four crystals

At the Hot Chips 2020 conference, Chief Architect and Senior Vice President of Intel’s Discrete Graphics Division, Raja Koduri, spoke about the theoretical performance of the company’s future accelerators. And the announced figures are impressive.

During his speech, Mr. Koduri took out the largest and most powerful GPU X from his pocket.e HP, which combines four crystals on one substrate (there will be simpler versions with one and two crystals). Each of them includes 512 computing units (EU, Execution Units) and, most likely, two HBM2e multilayer memory modules.

The supervisor then ran a test to show the excellent scalability of the X architecture.e and how increasing the number of dice on the substrate results in a multiple of the performance scaling. Working on a single crystal, the accelerator provides up to 10,588 teraflops of FP32 floating point at single precision. When two crystals are involved, performance scales almost perfectly to 21,161 teraflops (1.999x). The flagship quad-die solution delivers 3.993x the performance boost to 41.908 teraflops.

The distribution of tasks between GPUs in technologies such as SLI and CrossFire is much more difficult, and scaling from additional accelerators usually gives players a gain in the region of 50-80% at best. However, in compute workloads, tasks are often independent and can scale perfectly. Therefore, in professional tasks, Intel’s flagship GPU with four crystals will clearly show itself well. And here is how things will be at Intel with graphics scaling – independent tests will show.

For comparison: the GeForce RTX 2080 Ti graphics card in FP32 operations is capable of providing up to 14.2 teraflops of performance. But that’s not all: Mr. Koduri also mentioned that Xe HP on the basis of four crystals is capable of delivering performance in petaflops. In other words, using tensor blocks, Intel’s new GPU will be incredibly fast in machine learning and artificial intelligence tasks where high precision is not required.

If you notice an error, select it with the mouse and press CTRL + ENTER.

Leave a Comment