![]() |
Ginkgo Generated from branch based on master. Ginkgo version 1.8.0
A numerical linear algebra library targeting many-core architectures
|
This class describes the value and index types to be used when building a Ginkgo type from a configuration file. More...
#include <ginkgo/core/config/type_descriptor.hpp>
Public Member Functions | |
type_descriptor (std::string value_typestr="float64", std::string index_typestr="int32") | |
type_descriptor constructor. | |
const std::string & | get_value_typestr () const |
Get the value type string. | |
const std::string & | get_index_typestr () const |
Get the index type string. | |
This class describes the value and index types to be used when building a Ginkgo type from a configuration file.
A type_descriptor is passed in order to define the parse function defines which template parameters, in terms of value_type and/or index_type, the created object will have. For example, a CG solver created like this:
will have the value type float64
and the index type int32
. Any Ginkgo object that does not require one of these types will just ignore it. The value void
can be used to specify that no default type is provided. In this case, the configuration has to provide the necessary template types.
If the configuration specifies one of the fields (or both):
these types will take precedence over the type_descriptor.
|
explicit |
type_descriptor constructor.
There is free function make_type_descriptor
to create the object by template.
value_typestr | the value type string. "void" means no default. |
index_typestr | the index type string. "void" means no default. |