MicroBenchmark.cpp File Reference

#include "shambase/exception.hpp"
#include "shambase/stacktrace.hpp"
#include "shambase/string.hpp"
#include "shambase/time.hpp"
#include "sham/format/human_readable.hpp"
#include "shamalgs/collective/exchanges.hpp"
#include "shamalgs/collective/reduction.hpp"
#include "shambackends/Device.hpp"
#include "shambackends/benchmarks/fma_chains.hpp"
#include "shambackends/benchmarks/saxpy.hpp"
#include "shambackends/comm/CommunicationBuffer.hpp"
#include "shambackends/math.hpp"
#include "shamcomm/wrapper.hpp"
#include "shamsys/MicroBenchmark.hpp"
#include "shamsys/MpiWrapper.hpp"
#include "shamsys/NodeInstance.hpp"
#include "shamsys/legacy/log.hpp"
#include <stdexcept>
#include <vector>

Include dependency graph for MicroBenchmark.cpp:

Go to the source code of this file.

Namespaces
namespace	shamsys
	namespace for the system handling

Functions
void	shamsys::microbench::p2p_bandwidth (u32 wr_sender, u32 wr_receiv)
	MPI point-to-point bandwidth benchmark.
void	shamsys::microbench::p2p_latency (u32 wr1, u32 wr2)
	MPI point-to-point latency benchmark.
template<typename T>
void	shamsys::microbench::saxpy ()
	SAXPY benchmark, to get the maximum bandwidth.
template<typename T>
void	shamsys::microbench::fma_chains_rotation ()
	FMA chains benchmark to get the maximum floating point performance.
void	shamsys::microbench::vector_allgather (u32 el_per_rank)
	Vector allgather benchmark.

Detailed Description

Author

Timothée David–Cléris (tim.s.nosp@m.hamr.nosp@m.ock@p.nosp@m.roto.nosp@m.n.me)

Definition in file MicroBenchmark.cpp.

Function Documentation

fma_chains_rotation()

template<typename T>

void shamsys::microbench::fma_chains_rotation

(

)

FMA chains benchmark to get the maximum floating point performance.

Definition at line 343 of file MicroBenchmark.cpp.

Here is the call graph for this function:

p2p_bandwidth()

void shamsys::microbench::p2p_bandwidth	(	u32	wr_sender,
		u32	wr_receiv )

MPI point-to-point bandwidth benchmark.

Definition at line 100 of file MicroBenchmark.cpp.

Here is the call graph for this function:

p2p_latency()

void shamsys::microbench::p2p_latency	(	u32	wr1,
		u32	wr2 )

MPI point-to-point latency benchmark.

Definition at line 165 of file MicroBenchmark.cpp.

Here is the call graph for this function:

saxpy()

template<typename T>

void shamsys::microbench::saxpy

(

)

SAXPY benchmark, to get the maximum bandwidth.

Definition at line 230 of file MicroBenchmark.cpp.

Here is the call graph for this function:

vector_allgather()

void shamsys::microbench::vector_allgather

(

u32

el_per_rank

)

Vector allgather benchmark.

Definition at line 403 of file MicroBenchmark.cpp.

Here is the call graph for this function: