Vector Neural Network Instructions (VNNI)
GPTKB entity
Statements (23)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
CPU instruction set extension
|
| gptkbp:alternativeName |
AVX-VNNI
AVX512_VNNI |
| gptkbp:announced |
2018
|
| gptkbp:category |
SIMD instruction set
|
| gptkbp:feature |
fused multiply-add operations
support for INT8 and INT16 data types |
| https://www.w3.org/2000/01/rdf-schema#label |
Vector Neural Network Instructions (VNNI)
|
| gptkbp:introduced |
gptkb:Intel
|
| gptkbp:introducedIn |
gptkb:Intel_AVX-512
Intel AVX10 |
| gptkbp:purpose |
accelerate deep learning workloads
improve performance of convolutional neural networks |
| gptkbp:relatedTo |
gptkb:Intel_Deep_Learning_Boost
gptkb:AMX_(Advanced_Matrix_Extensions) |
| gptkbp:supportedBy |
gptkb:Intel_Cascade_Lake
gptkb:x86-64_architecture |
| gptkbp:target |
gptkb:machine_learning
AI inference acceleration |
| gptkbp:usedIn |
gptkb:Intel_Xeon_Scalable_processors
Intel Core processors (12th Gen and later) |
| gptkbp:bfsParent |
gptkb:Intel_DL_Boost
|
| gptkbp:bfsLayer |
7
|