aie::lut< ParallelAccesses, OffsetType, SlopeType > Struct Template Reference

Detailed Description

template<unsigned ParallelAccesses, typename OffsetType, typename SlopeType = OffsetType>
struct aie::lut< ParallelAccesses, OffsetType, SlopeType >

Abstraction to represent a LUT that is stored in memory, instantiated with pointer(s) to the already populated memory and the number of elements. To achieve a given level of access parallelism, the LUT values need to have a specific layout in memory:

• To achieve the 2 loads in parallel, the LUT needs to have 2 copies of the desired LUT values with repetition every 128b. For example with 32b values, in memory we would have the first 4 values (128b), then the same 4 again, then the next 4, which then repeat, etc.
• For 4 loads in parallel, we require the same layout as above, but two distinct copies like this, each placed in different banks. This is in the end 4 total copies of the LUT values in memory.
• For a single load without parallelism, the values are just stored linearly in memory.

Currently the only supported implementation is 4 parallel accesses on this architecture.

Template Parameters

ParallelAccesses	Defines how many parallel accesses will be done in a single LUT access, possibilities depend on the hardware available for the given architecture
OffsetType	Type of values stored within the lookup table.
SlopeType	Optional template parameter, only needed in certain cases of linear approximation where the offset/slope value pair uses two different types.

#include <aie.hpp>

Inheritance diagram for aie::lut< ParallelAccesses, OffsetType, SlopeType >:

Public Types
using	lut_impl = detail::lut< ParallelAccesses, OffsetType, SlopeType >
using	offset_type = OffsetType
using	slope_type = SlopeType

Public Member Functions
	lut (unsigned LUT_elems, const void *LUT_a)
	More...
	lut (unsigned LUT_elems, const void *LUT_ab)
	More...
	lut (unsigned LUT_elems, const void *LUT_ab, const void *LUT_cd)
	More...

Constructor & Destructor Documentation

lut() [1/3]

template<unsigned ParallelAccesses, typename OffsetType , typename SlopeType = OffsetType>

aie::lut< ParallelAccesses, OffsetType, SlopeType >::lut ( unsigned  LUT_elems, const void *  LUT_ab, const void *  LUT_cd  )

inline

Constructor for 4 parallel accesses. Each pointer points to an equivalent LUT populated within which the values are repeated twice, interleaved at a 128b granularity. In total the same values need to be present 4 times in memory to allow for the 4 parallel accesses.

Parameters

LUT_elems	Number elements in the LUT (not accounting for repetition).
LUT_ab	First two copies of the data, with the values repeated and interleaved at a 128b granularity.
LUT_cd	Next two copies of the data, with the values repeated and interleaved at a 128b granularity.

lut() [2/3]

template<unsigned ParallelAccesses, typename OffsetType , typename SlopeType = OffsetType>

aie::lut< ParallelAccesses, OffsetType, SlopeType >::lut ( unsigned  LUT_elems, const void *  LUT_ab  )

inline

Constructor for two parallel accesses.

Parameters

LUT_elems	Number of elements in the LUT (not accounting for repetition).
LUT_ab	Two copies of the data, with the values interleaved at a 128b granularity.

lut() [3/3]

template<unsigned ParallelAccesses, typename OffsetType , typename SlopeType = OffsetType>

aie::lut< ParallelAccesses, OffsetType, SlopeType >::lut ( unsigned  LUT_elems, const void *  LUT_a  )

inline

Constructor for singular access.

Parameters

LUT_elems	Number of elements in the LUT.
LUT_a	Pointer to the LUT values.

---

The documentation for this struct was generated from the following file:

• aie.hpp