AI Engine-ML v2 Intrinsics User Guide  v2025.1
Loading...
Searching...
No Matches
Vector Data Types

Topics

 Brain Floating-Point Vector Types
 
 Complex Floating-Point Vector Types
 
 Complex Vector Types
 
 Floating-Point Vector Types
 
 Integer Vector Types
 
 MX Vector Types
 Microscaling Formats (MX) vector data-types are stored in the vector registers within the AIE-ML v2.
 
 Sparse Vector Types
 

Detailed Description

Vector data-types pack multiple scalar data elements into a wider vector. There are vector data-types with standard sized elements (4-bit, 8-bit, 16-bit, 32-bit).

The general naming convention for these data-types is shown below.

    v{NumLanes}[cu]int{SizeofElement}
    v{NumLanes}bfloat16
    v{NumLanes}float[8,16]
    v{NumLanes}mx{6,9}

Disambiguation :

For example, v32uint4, v16int8, v8int16, v128int4, v32bfloat16, v32float16

The total data-width of the vector data-types can be 128-bits, 256-bits, 512-bits, or 1024-bits.

Note
Vector data-types are stored in different sets of registers within the processor core. Large vectors (>256-bits) occupy more than one physical register. Register spills will reduce performance when more than one large register is 'live'.
Supported vector types and sizes
Scalar typeSupported vector lanes
int4/uint432/64/128/256
int8/uint816/32/64/128
int16/uint168/16/32/64
int32/uint324/8/16/32
cint164/8/16/32
cint322/4/8/16
cfloat2/4/8/16/32
bfloat88/16/32/64/128/256
bfloat168/16/32/64/128
cbfloat164/8/16/32/64
float82/4/8/16/32/64/128/
float162/4/8/16/32/64/128
float2/4/8/16/32/64
mx416/64/128/256
mx616/64/128/256
mx916/32/64/128/256
int4/uint4_sparse64/128
int8/uint8_sparse128/256
int16/uint16_sparse64/128
float8_sparse128/256
float16_sparse64/128
bfloat8_sparse128/256
bfloat16_sparse64/128