Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
gaoqiong
MIGraphX
Commits
6ad2af4e
Commit
6ad2af4e
authored
Oct 08, 2022
by
Paul
Browse files
Handle transposes and data types
parent
c343c534
Changes
3
Hide whitespace changes
Inline
Side-by-side
Showing
3 changed files
with
77 additions
and
66 deletions
+77
-66
src/targets/gpu/fuse_ck.cpp
src/targets/gpu/fuse_ck.cpp
+12
-1
src/targets/gpu/jit/ck_gemm.cpp
src/targets/gpu/jit/ck_gemm.cpp
+44
-65
src/targets/gpu/jit/ck_gemm_instances.hpp
src/targets/gpu/jit/ck_gemm_instances.hpp
+21
-0
No files found.
src/targets/gpu/fuse_ck.cpp
View file @
6ad2af4e
...
@@ -22,9 +22,16 @@ struct ck_gemm
...
@@ -22,9 +22,16 @@ struct ck_gemm
}
}
std
::
string
name
()
const
{
return
"gpu::ck_gemm"
;
}
std
::
string
name
()
const
{
return
"gpu::ck_gemm"
;
}
void
check_gemm_shape
(
const
shape
&
s
)
const
{
if
(
contains
(
s
.
lens
(),
1
))
MIGRAPHX_THROW
(
"Invalid shape for ck_gemm"
);
}
shape
compute_shape
(
std
::
vector
<
shape
>
inputs
,
const
std
::
vector
<
module_ref
>&
mods
)
const
shape
compute_shape
(
std
::
vector
<
shape
>
inputs
,
const
std
::
vector
<
module_ref
>&
mods
)
const
{
{
check_shapes
{
inputs
,
*
this
}.
standar
d
();
check_shapes
{
inputs
,
*
this
}.
not_broadcaste
d
();
// if(mods.size() != 1)
// if(mods.size() != 1)
// MIGRAPHX_THROW("should have one submodule.");
// MIGRAPHX_THROW("should have one submodule.");
if
(
inputs
.
size
()
<
2
)
if
(
inputs
.
size
()
<
2
)
...
@@ -32,6 +39,8 @@ struct ck_gemm
...
@@ -32,6 +39,8 @@ struct ck_gemm
auto
n
=
inputs
.
size
();
auto
n
=
inputs
.
size
();
auto
a
=
inputs
[
n
-
2
];
auto
a
=
inputs
[
n
-
2
];
auto
b
=
inputs
[
n
-
1
];
auto
b
=
inputs
[
n
-
1
];
check_gemm_shape
(
a
);
check_gemm_shape
(
b
);
return
op
.
compute_shape
({
a
,
b
});
return
op
.
compute_shape
({
a
,
b
});
}
}
};
};
...
@@ -45,6 +54,8 @@ MIGRAPHX_PRED_MATCHER(is_ck_gemm, instruction_ref ins)
...
@@ -45,6 +54,8 @@ MIGRAPHX_PRED_MATCHER(is_ck_gemm, instruction_ref ins)
return
false
;
return
false
;
auto
a
=
ins
->
inputs
().
front
()
->
get_shape
();
auto
a
=
ins
->
inputs
().
front
()
->
get_shape
();
auto
b
=
ins
->
inputs
().
back
()
->
get_shape
();
auto
b
=
ins
->
inputs
().
back
()
->
get_shape
();
if
(
a
.
lens
().
size
()
>
2
or
b
.
lens
().
size
()
>
2
)
return
false
;
return
(
a
.
lens
()[
0
]
%
8
==
0
and
a
.
lens
()[
1
]
%
8
==
0
and
b
.
lens
()[
0
]
%
8
==
0
and
return
(
a
.
lens
()[
0
]
%
8
==
0
and
a
.
lens
()[
1
]
%
8
==
0
and
b
.
lens
()[
0
]
%
8
==
0
and
b
.
lens
()[
1
]
%
8
==
0
);
b
.
lens
()[
1
]
%
8
==
0
);
}
}
...
...
src/targets/gpu/jit/ck_gemm.cpp
View file @
6ad2af4e
...
@@ -38,6 +38,8 @@
...
@@ -38,6 +38,8 @@
#include <migraphx/pass_manager.hpp>
#include <migraphx/pass_manager.hpp>
#include <migraphx/env.hpp>
#include <migraphx/env.hpp>
#include "ck_gemm_instances.hpp"
namespace
migraphx
{
namespace
migraphx
{
inline
namespace
MIGRAPHX_INLINE_NS
{
inline
namespace
MIGRAPHX_INLINE_NS
{
...
@@ -77,93 +79,70 @@ __global__ void ck_gemm_kernel(void* a_p, void* b_p, void* c_p)
...
@@ -77,93 +79,70 @@ __global__ void ck_gemm_kernel(void* a_p, void* b_p, void* c_p)
)__migraphx__"
;
)__migraphx__"
;
std
::
size_t
int_div_ceil
(
std
::
size_t
x
,
std
::
size_t
y
)
{
return
(
x
+
y
-
1
)
/
y
;
}
static
std
::
size_t
int_div_ceil
(
std
::
size_t
x
,
std
::
size_t
y
)
{
return
(
x
+
y
-
1
)
/
y
;
}
static
std
::
size_t
block_size_index
=
13
;
std
::
size_t
get_
grid
_size
(
std
::
size_t
m
,
std
::
size_t
mpb
,
std
::
size_t
n
,
std
::
size_t
npb
)
static
std
::
size_t
get_
block
_size
(
const
std
::
vector
<
std
::
string
>&
s
)
{
{
return
int_div_ceil
(
m
,
mpb
)
*
int_div_ceil
(
n
,
npb
);
return
std
::
stoull
(
s
[
block_size_index
]
);
}
}
st
ruct
block_settings
st
atic
std
::
size_t
get_grid_size
(
const
std
::
vector
<
std
::
string
>&
s
,
std
::
size_t
m
,
std
::
size_t
n
)
{
{
int
bs
;
auto
mpb
=
std
::
stoull
(
s
[
block_size_index
+
1
]);
int
mpb
;
auto
npb
=
std
::
stoull
(
s
[
block_size_index
+
2
]);
int
npb
;
return
int_div_ceil
(
m
,
mpb
)
*
int_div_ceil
(
n
,
npb
);
};
}
namespace
fs
=
std
::
filesystem
;
struct
ck_gemm_compiler
:
compiler
<
ck_gemm_compiler
>
struct
ck_gemm_compiler
:
compiler
<
ck_gemm_compiler
>
{
{
// clang-format off
static
std
::
string
get_layout
(
const
shape
&
s
)
const
std
::
vector
<
std
::
string
>
instances
{
return
s
.
transposed
()
?
"ck::tensor_layout::gemm::ColumnMajor"
:
"ck::tensor_layout::gemm::RowMajor"
;
}
static
std
::
string
get_type
(
const
shape
&
s
)
{
{
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 256, 128, 32, 8, 2, 32, 32, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 32, 1, 8>, 8"
,
if
(
s
.
type
()
==
shape
::
half_type
)
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 256, 128, 32, 8, 8, 32, 32, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 32, 1, 8>, 8"
,
return
"ck::half_t"
;
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 256, 32, 8, 2, 32, 32, 2, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 32, 1, 8>, 8"
,
return
shape
::
cpp_type
(
s
.
type
());
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 256, 32, 8, 8, 32, 32, 2, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, 1, 1, 1, S<1, 32, 1, 8>, 8"
,
}
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 32, 8, 2, 32, 32, 2, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 32, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 32, 8, 8, 32, 32, 2, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 32, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 64, 32, 8, 2, 32, 32, 2, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<16,16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 32, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 64, 32, 8, 8, 32, 32, 2, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, 1, 1, 1, S<1, 32, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 64, 128, 32, 8, 2, 32, 32, 1, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 32, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 256, 64, 128, 32, 8, 8, 32, 32, 1, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 32, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 128, 128, 128, 32, 8, 2, 32, 32, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 16, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 128, 128, 128, 32, 8, 8, 32, 32, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, 1, 1, 1, S<1, 16, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 128, 128, 64, 32, 8, 2, 32, 32, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<8, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 32, 1, 4>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 128, 128, 64, 32, 8, 8, 32, 32, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 32, 1, 4>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 128, 64, 128, 32, 8, 2, 32, 32, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, 0, 1, 1, S<1, 16, 1, 8>, 8"
,
"Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmDefault, 1, 128, 64, 128, 32, 8, 8, 32, 32, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, 1, 1, 1, S<1, 16, 1, 8>, 8"
};
// clang-format on
const
std
::
vector
<
block_settings
>
params
{{
256
,
256
,
128
},
{
256
,
256
,
128
},
{
256
,
128
,
256
},
{
256
,
128
,
256
},
{
256
,
128
,
128
},
{
256
,
128
,
128
},
{
256
,
128
,
64
},
{
256
,
128
,
64
},
{
256
,
64
,
128
},
{
256
,
64
,
128
},
{
128
,
128
,
128
},
{
128
,
128
,
128
},
{
128
,
128
,
64
},
{
128
,
128
,
64
},
{
128
,
64
,
128
},
{
128
,
64
,
128
}};
std
::
vector
<
std
::
string
>
names
()
const
{
return
{
"ck_gemm"
,
"gpu::ck_gemm"
};
}
std
::
vector
<
std
::
string
>
names
()
const
{
return
{
"ck_gemm"
,
"gpu::ck_gemm"
};
}
operation
compile_op
(
context
&
/* ctx */
,
const
std
::
vector
<
shape
>&
inputs
,
const
value
&
v
)
const
operation
compile_op
(
context
&
/* ctx */
,
const
std
::
vector
<
shape
>&
inputs
,
const
value
&
v
)
const
{
{
int
i
=
v
.
get
(
"tuning_val"
,
4
);
auto
a_shape
=
inputs
[
0
];
assert
(
i
>=
0
and
i
<
instances
.
size
());
auto
b_shape
=
inputs
[
1
];
auto
c_shape
=
inputs
[
2
];
hip_compile_options
options
;
auto
m
=
c_shape
.
lens
().
front
();
auto
out_s
=
inputs
.
back
();
auto
n
=
c_shape
.
lens
().
back
();
auto
k
=
a_shape
.
lens
().
back
();
auto
sa
=
a_shape
.
strides
().
front
();
auto
sb
=
b_shape
.
strides
().
front
();
auto
sc
=
c_shape
.
strides
().
front
();
auto
b_s
=
params
[
i
];
int
i
=
v
.
get
(
"tuning_val"
,
4
);
auto
block_size
=
b_s
.
bs
;
const
auto
&
instance
=
get_instance
(
i
,
[
&
](
const
auto
&
x
)
->
bool
{
auto
m_per_block
=
b_s
.
mpb
;
return
get_layout
(
a_shape
)
==
x
[
0
]
and
auto
n_per_block
=
b_s
.
npb
;
get_layout
(
b_shape
)
==
x
[
1
]
and
auto
m
=
out_s
.
lens
().
front
();
get_layout
(
c_shape
)
==
x
[
2
]
and
auto
n
=
out_s
.
lens
().
back
();
get_type
(
a_shape
)
==
x
[
3
]
and
auto
grid_size
=
get_grid_size
(
m
,
m_per_block
,
n
,
n_per_block
);
get_type
(
b_shape
)
==
x
[
4
]
and
get_type
(
c_shape
)
==
x
[
5
];
});
options
.
set_launch_params
(
v
,
grid_size
*
block_size
,
block_size
);
hip_compile_options
options
;
options
.
set_launch_params
(
v
,
get_grid_size
(
instance
,
m
,
n
),
get_block_size
(
instance
));
options
.
inputs
=
inputs
;
options
.
inputs
=
inputs
;
options
.
output
=
out_s
;
options
.
output
=
c_shape
;
options
.
kernel_name
=
"ck_gemm_kernel"
;
options
.
kernel_name
=
"ck_gemm_kernel"
;
options
.
virtual_inputs
=
inputs
;
options
.
virtual_inputs
=
inputs
;
auto
k
=
inputs
.
front
().
lens
().
back
();
auto
sa
=
inputs
.
front
().
strides
().
front
();
auto
sb
=
inputs
.
at
(
1
).
strides
().
front
();
auto
sc
=
inputs
.
back
().
strides
().
front
();
auto
src
=
interpolate_string
(
ck_gemm_kernel
,
auto
src
=
interpolate_string
(
ck_gemm_kernel
,
{{
"instance"
,
instance
s
[
i
]
},
{{
"instance"
,
join_strings
(
instance
,
","
)
},
{
"m"
,
to_string
(
m
)},
{
"m"
,
to_string
(
m
)},
{
"k"
,
to_string
(
k
)},
{
"k"
,
to_string
(
k
)},
{
"n"
,
to_string
(
n
)},
{
"n"
,
to_string
(
n
)},
...
...
src/targets/gpu/jit/ck_gemm_instances.hpp
0 → 100644
View file @
6ad2af4e
#ifndef MIGRAPHX_GUARD_JIT_CK_INSTANCES_HPP
#define MIGRAPHX_GUARD_JIT_CK_INSTANCES_HPP
#include <algorithm>
#include <vector>
#include <string>
#include <functional>
inline
const
std
::
vector
<
std
::
string
>&
get_instance
(
std
::
size_t
i
,
const
std
::
function
<
bool
(
const
std
::
vector
<
std
::
string
>&
)
>&
pred
)
{
static
std
::
vector
<
std
::
vector
<
std
::
vector
<
std
::
string
>>>
instances
=
{{{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"1"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
}},{{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"64"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"32"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"32"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"64"
,
"32"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"32"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,8>"
,
"4"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"256"
,
"256"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"256"
,
"128"
,
"256"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"128"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"256"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"128"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"128"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"64"
,
"64"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"256"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"256"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"128"
,
"128"
,
"32"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"128"
,
"32"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"64"
,
"64"
,
"32"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"2"
,
"64"
,
"32"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"8"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"16"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
}},{{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"4"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,64,1,4>"
,
"16"
}},{{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"1"
,
"1"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
}},{{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"2"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"64"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"32"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"32"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"64"
,
"32"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"32"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"8"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"64"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"32"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"32"
,
"128"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,4>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"64"
,
"32"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,2>"
,
"16"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"int8_t"
,
"int8_t"
,
"int8_t"
,
"int32_t"
,
"int32_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"64"
,
"32"
,
"64"
,
"64"
,
"16"
,
"16"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"16"
,
"16"
,
"1"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,2>"
,
"16"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,8,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"4"
,
"1"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"1"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"float"
,
"float"
,
"float"
,
"float"
,
"float"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"16"
,
"4"
,
"4"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"4"
,
"4"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"4"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,16>"
,
"4"
}},{{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"2"
,
"2"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::ColumnMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
}},{{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"256"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"256"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"4"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"4"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<8,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,4>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"128"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"2"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,16,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<16,16,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"128"
,
"64"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"2"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"1"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"8"
,
"2"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<8,32,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"4"
,
"2"
,
"0"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
},{
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::tensor_layout::gemm::RowMajor"
,
"ck::half_t"
,
"ck::half_t"
,
"ck::half_t"
,
"float"
,
"ck::half_t"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::element_wise::PassThrough"
,
"ck::tensor_operation::device::GemmSpecialization::Default"
,
"1"
,
"256"
,
"64"
,
"128"
,
"32"
,
"8"
,
"8"
,
"32"
,
"32"
,
"1"
,
"2"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<1,0,2>"
,
"ck::Sequence<1,0,2>"
,
"2"
,
"8"
,
"8"
,
"1"
,
"ck::Sequence<4,64,1>"
,
"ck::Sequence<0,2,1>"
,
"ck::Sequence<0,2,1>"
,
"1"
,
"2"
,
"8"
,
"1"
,
"1"
,
"1"
,
"ck::Sequence<1,32,1,8>"
,
"8"
}}};
auto
it
=
std
::
find_if
(
instances
.
begin
(),
instances
.
end
(),
[
&
](
const
auto
&
v
)
{
return
pred
(
v
[
0
]);
});
return
it
->
at
(
i
);
}
#endif
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment