Advanced Matrix Extensions (AMX), also known as Intel Advanced Matrix Extensions (Intel AMX), are extensions to the x86 instruction set architecture (ISA) for microprocessors from Intel designed to work on matrices to accelerate artificial intelligence (AI) and machine learning (ML) workloads. They perform matrix multiplication at the hardware level, making them apt for problems and algorithms that use matrix multiplication as their core.
16×J for
A and
J is 64 for INT8 and 32 for BF16. The matrix multiplication requires 512J operations in 16 cycles.
- Compiler and assembler support
- Operating system support
- glibc support for detecting AMX feature in CPUs committed on 25 June 2020
- Linux kernel support since version 5.16
- VMware vSphere support for AMX in virtual machines released in ESXi version 8.0u1 for VMs using Hardware Version 20