Accumulator Element Types

Accumulators in AIE API rely on the aie::accum class template. More...

Overview

Accumulators in AIE API rely on the aie::accum class template.

The first argument of this template is an element type tag that specifies the accumulation class (integer, complex, floating point, ...) and the required accumulation bits.

Classes
struct	acc32
	Tag used to request an accumulator with at least 32 bit per element. More...
struct	acc40
	Tag used to request an accumulator with at least 40 bit per element. More...
struct	acc48
	Tag used to request an accumulator with at least 48 bit per element. More...
struct	acc56
	Tag used to request an accumulator with at least 56 bit per element. More...
struct	acc64
	Tag used to request an accumulator with at least 64 bit per element. More...
struct	acc72
	Tag used to request an accumulator with at least 72 bit per element. More...
struct	acc80
	Tag used to request an accumulator with at least 80 bit per element. More...
struct	accauto
	Internal tag used to signal that the default accumulator precision is needed. More...
struct	accfloat
	Tag used to request an accumulator with single precision floating point elements. More...
struct	cacc32
	Tag used to request an accumulator with complex elements of at least 32 bit per component. More...
struct	cacc40
	Tag used to request an accumulator with complex elements of at least 40 bit per component. More...
struct	cacc48
	Tag used to request an accumulator with complex elements of at least 48 bit per component. More...
struct	cacc56
	Tag used to request an accumulator with complex elements of at least 56 bit per component. More...
struct	cacc64
	Tag used to request an accumulator with complex elements of at least 64 bit per component. More...
struct	cacc72
	Tag used to request an accumulator with complex elements of at least 72 bit per component. More...
struct	cacc80
	Tag used to request an accumulator with complex elements of at least 80 bit per component. More...
struct	caccfloat
	Tag used to request an accumulator with complex elements of single precision floating point components. More...

---

Class Documentation

acc32

struct acc32

Tag used to request an accumulator with at least 32 bit per element.

acc40

struct acc40

Tag used to request an accumulator with at least 40 bit per element.

acc48

struct acc48

Tag used to request an accumulator with at least 48 bit per element.

acc56

struct acc56

Tag used to request an accumulator with at least 56 bit per element.

acc64

struct acc64

Tag used to request an accumulator with at least 64 bit per element.

acc72

struct acc72

Tag used to request an accumulator with at least 72 bit per element.

Note

Supported in AIE only.

acc80

struct acc80

Tag used to request an accumulator with at least 80 bit per element.

Note

Supported in AIE only.

accauto

struct accauto

Internal tag used to signal that the default accumulator precision is needed.

This type is not meant to be directly used by AIE API users.

accfloat

struct accfloat

Tag used to request an accumulator with single precision floating point elements.

cacc32

struct cacc32

Tag used to request an accumulator with complex elements of at least 32 bit per component.

cacc40

struct cacc40

Tag used to request an accumulator with complex elements of at least 40 bit per component.

cacc48

struct cacc48

Tag used to request an accumulator with complex elements of at least 48 bit per component.

cacc56

struct cacc56

Tag used to request an accumulator with complex elements of at least 56 bit per component.

cacc64

struct cacc64

Tag used to request an accumulator with complex elements of at least 64 bit per component.

cacc72

struct cacc72

Tag used to request an accumulator with complex elements of at least 72 bit per component.

Note

Supported in AIE only.

cacc80

struct cacc80

Tag used to request an accumulator with complex elements of at least 80 bit per component.

Note

Supported in AIE only.

caccfloat

struct caccfloat

Tag used to request an accumulator with complex elements of single precision floating point components.