Matrix (AMX and ACE)
AMX-FP16
Tile matrix multiplication of IEEE half-precision values accumulated into single precision.
- CPUID
- AMX_FP16 (leaf 1EH)
CPUID.(EAX=1EH,ECX=1):EAX[3]or AMX_TILE + AMX_FP16CPUID.(EAX=07H,ECX=0):EDX[24]CPUID.(EAX=07H,ECX=1):EAX[21] - Instructions
- 1 mnemonic
- Processors
- Granite Rapids, Diamond Rapids
AMX-FP16 adds TDPFP16PS, which multiplies tiles of IEEE half-precision pairs and accumulates in single precision, the FP16 counterpart of TDPBF16PS. Intel introduced it with Granite Rapids.
Instructions
| Mnemonic | Summary | Forms |
|---|---|---|
| TDPFP16PS | Multiplies two tiles of half-precision pairs and accumulates each row-by-column dot product into a single-precision tile. | 1 |
CPUID and processors
XED splits AMX-FP16 into 1 ISA set. Software can use a form when the processor reports every CPUID bit of one alternative of the form's ISA set.
| ISA set | CPUID | Processors in XED |
|---|---|---|
AMX_FP16 | AMX_FP16 (leaf 1EH) or AMX_TILE + AMX_FP16 | Granite Rapids, Diamond Rapids |
Sources
- Intel XED v2026.08.23 (commit
0bcb6237345c): forms, encodings, ISA sets, CPUID bits and chip model. - Intel 64 and IA-32 Architectures Software Developer's Manual, Volume 2 (325383-092, June 2026)
- Intel Architecture Instruction Set Extensions and Future Features Programming Reference (319433-062, June 2026)
The tables are derived from Intel XED, Copyright Intel Corporation, licensed under the Apache License 2.0; x86-64.net converted and reformatted them. The text is our own.