Overview

xmat is a cross platform serialization library meant for multi-dimensional numeric arrays structured in key-value format.

xmat provides:

  • binary format for file and network data exchange.

  • TCP-sockets aimed for this format

  • data types supported:

    • integer and float numeric types (including complex numbers)

    • ascii-characters

  • programming languages supported: C++14+, Python 3.10+, Matlab

  • the following C++ compilers are supported::

    • Unix: GCC

    • Windows: Visual C++, MinGW 12

Types supported

types summary

xmat

C++

Python

NumPy

Matlab

Size

Id

string

char

str

numpy.str_

char

1

0x01

[x] bool

bool

bool

numpy.bool_

logical

1

0x02

int8

std::int8_t

–

numpy.int8

int8

1

0x10

int16

std::int16_t

–

numpy.int16

int16

2

0x11

int32

std::int32_t

–

numpy.int32

int32

4

0x12

int64

std::int64_t

int

numpy.int64

int64

8

0x13

complex int8

std::complex<std::int8_t>

–

–

–

2

0x20

complex int16

std::complex<std::int16_t>

–

–

–

4

0x21

complex int32

std::complex<std::int32_t>

–

–

–

8

0x22

complex int64

std::complex<std::int64_t>

–

–

–

16

0x23

uint8

std::uint8_t

–

numpy.uint8

uint8

1

0x30

uint16

std::uint16_t

–

numpy.uint16

uint16

2

0x31

uint32

std::uint32_t

–

numpy.uint32

uint32

4

0x32

uint64

std::uint64_t

–

numpy.uint64

uint64

8

0x33

complex uint8

std::complex<std::uint8_t>

–

–

–

2

0x40

complex uint16

std::complex<std::uint16_t>

–

–

–

4

0x41

complex uint32

std::complex<std::uint32_t>

–

–

–

8

0x42

complex uint64

std::complex<std::uint64_t>

–

–

–

16

0x43

float32

float

–

numpy.float32

single

4

0x52

float64

double

float

numpy.float64

double

8

0x53

complex float32

std::complex<float>

–

numpy.complex64

single

8

0x62

complex float64

std::complex<double>

complex

numpy.complex128

double

16

0x63

Benchmark

benchmark

Size

1K

4K

16K

64K

256K

1024K

C++: char

1

4

16

64

256

1024

C++: double

1

4

16

64

256

1024

C++: complex double

1

4

16

64

256

1024

Python: char

1

4

16

64

256

1024

Python: double

1

4

16

64

256

1024

Python: complex double

1

4

16

64

256

1024

Matlab: char

1

4

16

64

256

1024

Matlab: double

1

4

16

64

256

1024

Matlab: complex double

1

4

16

64

256

1024