Accelerating AI and machine learning, Intel confirms that Arrow Lake-S and Lunar Lake support AVX-VNNI instruction set

Thanks to CTOnews.com netizens Wu Yanzu in South China for the delivery of clues! CTOnews.com, July 4, Intel recently upgraded its architecture instruction set extension and subsequent function documentation to adjust the specifications of next-generation processors, but also specifically mentioned that the Arrow Lake and Lunar Lake processor families will support instruction sets such as SHA512, SM3 / SM4 and AVX-VNNI.

Of particular concern is that the Arrow Lake and Lunar Lake processor families will support the AVX-VNNI instruction set, which improves performance when dealing with neural networks by providing specialized algorithms for 8-bit / 16-bit integer operations.

CTOnews.com cited the following supporting documents, Intel believes that the new instruction set can speed up the running of artificial intelligence, machine learning and deep learning algorithms applications, more effective performance of the processor.

The document also mentions that the above two consumer processors have added support for SHA512, SM3 and SM4 instructions, increasing their security and encryption capabilities.

SHA512 is a widely used encrypted hash function, and its integration ensures better data integrity and secure data transmission.

In addition, the file also stipulates that SM3 and SM4 encryption hash algorithms are used in data exchange to enhance security.