LLVM Compiler Adds Support For Next-Gen Intel Arrow Lake & Lunar Lake CPUs

5/5 - (10 votes)

The LLVM 18 compiler has recently added support for Intel’s upcoming Arrow Lake and Lunar Lake CPUs, which are scheduled to be released in 2024. This update comes after the GCC 14 compiler introduced several instruction sets, including AVX-VNNI-INT16, SHA512, SM3, and SM4, which are crucial for enhancing AI and machine learning applications in future CPUs, starting with Intel’s Meteor Lake lineup. According to Phoronix, LLVM has now integrated these instructions into its compiler, along with “-march= CPU” targets specific to Intel architecture, including Arrow Lake and Lunar Lake CPUs.

Additionally, the report mentions that the Intel Arrow Lake and Lunar Lake CPUs will continue to utilize the existing Alder Lake scheduler model and tuning. The aforementioned instructions are specifically applicable to the Arrow Lake-S and Lunar Lake variants. It is worth noting that the standard Arrow Lake chips for mobility are expected to have similar instructions to Sierra Forest chips, which will be released in early 2024. It is noteworthy that Intel Arrow Lake-S is explicitly listed, as these desktop parts will be compatible with the LGA 1851 socketed platform on the 800-series motherboards, expected to be available in late 2024. Conversely, Lunar Lake, like Meteor Lake, is expected to be exclusively launched for mobile devices.

These additions are significant as they provide insights into the development stages prior to the release of a new architecture. Moreover, small pieces of information can often hint at new features that will be included in future releases. It is commendable to see Intel’s efforts in providing preliminary support for its upcoming CPU lineup. In comparison, AMD is known for preparing such updates just before the official launch, sometimes as late as a month in advance.

In conclusion, the LLVM 18 compiler now supports Intel’s upcoming Arrow Lake and Lunar Lake CPUs. These CPUs are anticipated to be released in 2024 and will offer improved AI and machine learning capabilities. Intel’s decision to integrate these instructions into its compiler demonstrates their commitment to providing early support for their upcoming lineup.

About Benjamin Harris

Leave a Reply

Your email address will not be published. Required fields are marked *