]> Git Repo - qemu.git/commit
i386/cpu: Expose AVX_VNNI instruction to guest
authorYang Zhong <[email protected]>
Wed, 7 Apr 2021 01:56:09 +0000 (09:56 +0800)
committerPaolo Bonzini <[email protected]>
Wed, 26 May 2021 12:49:45 +0000 (14:49 +0200)
commitc1826ea6a052084f2e6a0bae9dd5932a727df039
treeeb488f76f291b20cd220ad1132ad54880d166273
parentcfa1f4bceea02c386bd7408fcc29544ac4679657
i386/cpu: Expose AVX_VNNI instruction to guest

Expose AVX (VEX-encoded) versions of the Vector Neural Network
Instructions to guest.

The bit definition:
CPUID.(EAX=7,ECX=1):EAX[bit 4] AVX_VNNI

The following instructions are available when this feature is
present in the guest.
  1. VPDPBUS: Multiply and Add Unsigned and Signed Bytes
  2. VPDPBUSDS: Multiply and Add Unsigned and Signed Bytes with Saturation
  3. VPDPWSSD: Multiply and Add Signed Word Integers
  4. VPDPWSSDS: Multiply and Add Signed Integers with Saturation

As for the kvm related code, please reference Linux commit id 1085a6b585d7.

The release document ref below link:
https://software.intel.com/content/www/us/en/develop/download/\
intel-architecture-instruction-set-extensions-programming-reference.html

Signed-off-by: Yang Zhong <[email protected]>
Message-Id: <20210407015609[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>
target/i386/cpu.c
target/i386/cpu.h
This page took 0.026472 seconds and 4 git commands to generate.