EvaluateImpResults Class Reference

Class with operator function to evaluate implementations of the same algorithm across different data types and acceleration methods. More...

#include <EvaluateImpResults.h>

Inheritance diagram for EvaluateImpResults:

Public Member Functions
std::pair< MultRunData, std::vector< RunSpeedupAvgMedian > >	EvalResultsSingDataTypeAcc (const MultRunData &run_results, const run_environment::RunImpSettings &run_imp_settings, size_t data_size) const
	Evaluate results for implementation runs on multiple inputs with all the runs having the same data type and acceleration method return run data results with run speedups added as well as average and median speedups with headers describing speedups. More...
void	EvalAllResultsWriteOutput (const std::unordered_map< size_t, MultRunDataWSpeedupByAcc > &run_results_mult_runs, const run_environment::RunImpSettings &run_imp_settings, run_environment::AccSetting opt_imp_acc) const
	Evaluate results for all implementation runs on multiple inputs with the runs potentially having different data types and acceleration methods and write run result and speedup outputs to files. More...

Detailed Description

Class with operator function to evaluate implementations of the same algorithm across different data types and acceleration methods.

Definition at line 57 of file EvaluateImpResults.h.

Member Function Documentation

EvalAllResultsWriteOutput()

void EvaluateImpResults::EvalAllResultsWriteOutput	(	const std::unordered_map< size_t, MultRunDataWSpeedupByAcc > &	run_results_mult_runs,
		const run_environment::RunImpSettings &	run_imp_settings,
		run_environment::AccSetting	opt_imp_acc
	)		const

Evaluate results for all implementation runs on multiple inputs with the runs potentially having different data types and acceleration methods and write run result and speedup outputs to files.

Parameters

run_results_mult_runs
run_imp_settings
opt_imp_acc

Definition at line 173 of file EvaluateImpResults.cpp.

EvalResultsSingDataTypeAcc()

std::pair< MultRunData, std::vector< RunSpeedupAvgMedian > > EvaluateImpResults::EvalResultsSingDataTypeAcc	(	const MultRunData &	run_results,
		const run_environment::RunImpSettings &	run_imp_settings,
		size_t	data_size
	)		const

Evaluate results for implementation runs on multiple inputs with all the runs having the same data type and acceleration method return run data results with run speedups added as well as average and median speedups with headers describing speedups.

Parameters

run_results
run_imp_settings
data_size

Returns

Run results with additional data and computed speedups from run

Definition at line 40 of file EvaluateImpResults.cpp.

---

The documentation for this class was generated from the following files:

• src/RunEval/EvaluateImpResults.h
• src/RunEval/EvaluateImpResults.cpp