GCC Compiler Patches Arrive For Intel Meteor Lake & Sierra Forest CPUs

Jason R. Wilson
GCC Compiler Patches Arrive For Intel Meteor Lake & Sierra Forest CPUs 1
Image source: J. Wilson, Wccftech.

Intel has published new patches for the GNU Compiler Collection or GCC, that will assist with support for upcoming Meteor Lake and Sierra Forest CPUs.

Meteor Lake and Sierra Forest from Intel receive GCC compiler patchwork before full implementation

The GCC includes several coding languages and libraries, such as C, C++, and more, to offer an open-source compiling option for users utilizing operating systems like Linux, Ubuntu, and others. The new compiler updates provided by Intel will assist with bringing in introductory x86_64 extensions for the instruction set seen in the Sierra Forest CPUs that will produce high core counts for Intel.

Related StoryJason R. Wilson
ASUS Reveals ExpertBook B9, Utilizing Intel’s 12th Gen Alder Lake vPro CPUs

Intel fellow and engineer Haochen was the first to release a Meteor Lake patch to the GCC compiler following a separate release for Raptor Lake. However, these two patchwork fixes were minimal as the changes were not highly integral to the overall performance of the two architectures.

The Sierra Forest, Intel's Xeon server series CPUs that rely strictly on E-cores, also saw inclusion in the newest update, where most of the work was more critical to improvements. First off, the "sierraforest" compiler target was enabled to offer further instructions to the next-gen server processors, such as:

  • AVX-IFMA
  • AVX-VNNI-INT8
  • AVX-NE-CONVERT
  • CMPccXADD

Michael Larabel of the website Phoronix recovered this list from the GNU Project site, along with this programming reference from Intel, to provide information about the new instructions.

Image capture the x86_64 ISA programming reference manual to show the release into Sierra Forest CPUs. Image source: Intel via Phoronix.

The top three instructions are focused on AVX-512 video codec, and the first covers the Integer Fused Multiply Add command, while the Vector Neural Network Instructions utilize the INT8 inference. AVX-NE-CONVERT installs the BF 16 FP element to the FP32 by converting it from m16. Lastly, CMPccXADD stands for "Compare and ADD if Condition is Met," the last set of instructions for AVX-512 for this GCC patch series.

Larabel notes that the new updates can still be integrated into the GCC 13 compiler, which is planned for release in 2023, following Ubuntu 23.10 later in the year. The timing will be perfect for Intel as the company will release the Sierra Forest processors for data center use towards the fourth quarter of 2023.

News Sources: Phoronix, GNU Compiler, Intel x86_64 ISA programming reference manual (PDF download),

Share this story

Deal of the Day

Comments