Intel Advanced Matrix Extensions
GPTKB entity
Statements (25)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:architecture
|
| gptkbp:abbreviation |
gptkb:AMX
|
| gptkbp:announced |
2020
|
| gptkbp:architecture |
x86
|
| gptkbp:category |
SIMD extension
|
| gptkbp:developer |
gptkb:Intel
|
| gptkbp:documentation |
gptkb:Intel_64_and_IA-32_Architectures_Software_Developer’s_Manual
|
| gptkbp:enables |
higher throughput for matrix operations
|
| gptkbp:feature |
tile registers
tile operations |
| gptkbp:firstSupportedCPU |
gptkb:Sapphire_Rapids
|
| https://www.w3.org/2000/01/rdf-schema#label |
Intel Advanced Matrix Extensions
|
| gptkbp:introducedIn |
gptkb:Sapphire_Rapids
|
| gptkbp:purpose |
accelerate matrix multiplication
|
| gptkbp:relatedTo |
gptkb:Intel_Deep_Learning_Boost
gptkb:Intel_AVX-512 |
| gptkbp:successor |
Intel AVX-512 VNNI
|
| gptkbp:supportedBy |
gptkb:Intel_Xeon_Scalable_(Sapphire_Rapids)
|
| gptkbp:tileRegisterCount |
8
|
| gptkbp:tileRegisterSize |
1024 bytes
|
| gptkbp:usedFor |
gptkb:machine_learning
high-performance computing artificial intelligence workloads |
| gptkbp:bfsParent |
gptkb:Sapphire_Rapids
|
| gptkbp:bfsLayer |
5
|