Prototype implementation to test C++/C# marshalling and the different matrix multiplication approaches. More...

Static Public Member Functions
static int	Benchmark (int nrCandidates, int nrSpectra, int topN, int batchSize, Random r)
	Function to run synthetic benchmarks of all the different matrix multiplication functions.

static int	Compare (int nrCandidates, int nrSpectra, int topN, int batchSize, Random r)
	Function to compare the results of all the different matrix multiplication functions.

static int	Cuda (int nrCandidates, int nrSpectra, int topN, int batchSize, Random r, bool batched, int batchMode)
	Wrapper for testing GPU-based matrix multiplication functions.

static void	Main (string[] args)
	Main function to be executed when calling the executable. The routine to be called depends on the commandline arguments passed to the function. The commandline call has to look like this: DataLoader.exe [(string)Mode][(int)NrCandidates][(int)NrSpectra][(int)TopN][(int)BatchSize].

static int	DeterministicCompare ()
	[DEPRECATED] Function for testing a simple matrix multiplication of a deterministic problem. Will not work with the current CandidateVectorSearch DLLs.

static int	Eigen (int nrCandidates, int nrSpectra, int topN, int batchSize, Random r, bool batched, bool sparse, bool useInt=false)
	Wrapper for testing CPU-based matrix multiplication functions.

Detailed Description

Prototype implementation to test C++/C# marshalling and the different matrix multiplication approaches.

Definition at line 6 of file Benchmark.cs.

Member Function Documentation

◆ Benchmark()

static int CandidateVectorSearch.DataLoader.Benchmark	(	int	nrCandidates,
		int	nrSpectra,
		int	topN,
		int	batchSize,
		Random	r
	)

static

Function to run synthetic benchmarks of all the different matrix multiplication functions.

Parameters

nrCandidates	The number of candidates that should be simulated.
nrSpectra	The number of spectra to be simulated.
topN	The number of top hits returned for every spectrum.
batchSize	The number of spectra processed at once for matrix * matrix approaches.
r	A random number generator used for simulation.

Returns: Returns 0 if the benchmarks finished successfully.

Definition at line 17 of file Benchmark.cs.

◆ Compare()

static int CandidateVectorSearch.DataLoader.Compare	(	int	nrCandidates,
		int	nrSpectra,
		int	topN,
		int	batchSize,
		Random	r
	)

static

Function to compare the results of all the different matrix multiplication functions.

Parameters

nrCandidates	The number of candidates that should be simulated.
nrSpectra	The number of spectra to be simulated.
topN	The number of top hits returned for every spectrum.
batchSize	The number of spectra processed at once for matrix * matrix approaches.
r	A random number generator used for simulation.

Returns: Returns 0 if the function finished successfully.

Definition at line 17 of file Compare.cs.

◆ Cuda()

static int CandidateVectorSearch.DataLoader.Cuda	(	int	nrCandidates,
		int	nrSpectra,
		int	topN,
		int	batchSize,
		Random	r,
		bool	batched,
		int	batchMode
	)

static

Wrapper for testing GPU-based matrix multiplication functions.

Parameters

nrCandidates	The number of candidates that should be simulated.
nrSpectra	The number of spectra to be simulated.
topN	The number of top hits returned for every spectrum.
batchSize	The number of spectra processed at once for matrix * matrix approaches.
r	A random number generator used for simulation.
batched	Whether to run a vector * matrix (false) or matrix * matrix (true) approach.
batchMode	Which matrix * matrix approach to use if batched = true. If batchMode = 2 uses sparse matrix * dense matrix approach, otherwise sparse matrix * sparse matrix approach is used.

Returns: Returns 0 if the function finished successfully.

Definition at line 52 of file Cuda.cs.

◆ DeterministicCompare()

static int CandidateVectorSearch.DataLoader.DeterministicCompare ( )

static

[DEPRECATED]
Function for testing a simple matrix multiplication of a deterministic problem.
Will not work with the current CandidateVectorSearch DLLs.

Returns: Returns 0 if the function finished successfully.

Definition at line 14 of file DeterministicCompare.cs.

◆ Eigen()

static int CandidateVectorSearch.DataLoader.Eigen	(	int	nrCandidates,
		int	nrSpectra,
		int	topN,
		int	batchSize,
		Random	r,
		bool	batched,
		bool	sparse,
		bool	useInt = `false`
	)

static

Wrapper for testing CPU-based matrix multiplication functions.

Parameters

nrCandidates	The number of candidates that should be simulated.
nrSpectra	The number of spectra to be simulated.
topN	The number of top hits returned for every spectrum.
batchSize	The number of spectra processed at once for matrix * matrix approaches.
r	A random number generator used for simulation.
batched	Whether to run a vector * matrix (false) or matrix * matrix (true) approach.
sparse	Whether to run a sparse * dense (false) or sparse * sparse (true) approach.
useInt	Whether or not to convert floating point numbers to integers before multiplication.

Returns: Returns 0 if the function finished successfully.

Definition at line 100 of file Eigen.cs.

◆ Main()

static void CandidateVectorSearch.DataLoader.Main ( string[] args )

static

Main function to be executed when calling the executable.
The routine to be called depends on the commandline arguments passed to the function.
The commandline call has to look like this:
DataLoader.exe [(string)Mode][(int)NrCandidates][(int)NrSpectra][(int)TopN][(int)BatchSize].

Parameters

args	Array of commandline arguments.

Definition at line 43 of file DataLoader.cs.

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

C:/Users/P42587/Documents/GitHub/CandidateVectorSearch/DataLoader/Benchmark.cs
C:/Users/P42587/Documents/GitHub/CandidateVectorSearch/DataLoader/Cuda.cs
C:/Users/P42587/Documents/GitHub/CandidateVectorSearch/DataLoader/DataLoader.cs
C:/Users/P42587/Documents/GitHub/CandidateVectorSearch/DataLoader/Eigen.cs
C:/Users/P42587/Documents/GitHub/CandidateVectorSearch/DataLoader/Compare.cs
C:/Users/P42587/Documents/GitHub/CandidateVectorSearch/DataLoader/DeterministicCompare.cs

Static Public Member Functions

Detailed Description

Member Function Documentation

◆ Benchmark()

◆ Compare()

◆ Cuda()

◆ DeterministicCompare()

◆ Eigen()

◆ Main()