quantize_contraction.h File Reference

#include <nlohmann/json_fwd.hpp>
#include <string>
#include "sdfg/analysis/analysis.h"
#include "sdfg/builder/structured_sdfg_builder.h"
#include "sdfg/data_flow/library_node.h"
#include "sdfg/transformations/transformation.h"

Go to the source code of this file.

Classes
class	sdfg::fp_math::QuantizeContraction
	Computes a floating-point matmul (MatMul or GEMM node) on integers, quantized with a scale per row of A and per column of B. More...

Namespaces
namespace	sdfg
namespace	sdfg::fp_math

Enumerations
enum class	sdfg::fp_math::Rescaling { sdfg::fp_math::Exponents , sdfg::fp_math::Direct }
	How QuantizeContraction applies the scales of A's rows and B's columns to the Int32 products. More...