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
composable_kernel_ROCM
Commits
dbe06911
Commit
dbe06911
authored
Sep 26, 2023
by
Umang Yadav
Browse files
Remove unnecessary changes
parent
bc731b90
Changes
14
Hide whitespace changes
Inline
Side-by-side
Showing
14 changed files
with
16 additions
and
36 deletions
+16
-36
include/ck/ck.hpp
include/ck/ck.hpp
+4
-5
include/ck/host_utility/device_prop.hpp
include/ck/host_utility/device_prop.hpp
+0
-1
include/ck/host_utility/hip_check_error.hpp
include/ck/host_utility/hip_check_error.hpp
+0
-1
include/ck/host_utility/io.hpp
include/ck/host_utility/io.hpp
+0
-1
include/ck/host_utility/stream_utility.hpp
include/ck/host_utility/stream_utility.hpp
+0
-1
include/ck/problem_transform/transform_forward_convolution3d_into_gemm_v4r4r4_ndhwc_kzyxc_ndhwk.hpp
...ward_convolution3d_into_gemm_v4r4r4_ndhwc_kzyxc_ndhwk.hpp
+0
-1
include/ck/tensor/static_tensor.hpp
include/ck/tensor/static_tensor.hpp
+0
-1
include/ck/tensor_description/cluster_descriptor.hpp
include/ck/tensor_description/cluster_descriptor.hpp
+0
-1
include/ck/tensor_description/multi_index_transform.hpp
include/ck/tensor_description/multi_index_transform.hpp
+0
-1
include/ck/tensor_description/multi_index_transform_helper.hpp
...de/ck/tensor_description/multi_index_transform_helper.hpp
+0
-1
include/ck/tensor_description/tensor_adaptor.hpp
include/ck/tensor_description/tensor_adaptor.hpp
+10
-15
include/ck/tensor_description/tensor_descriptor.hpp
include/ck/tensor_description/tensor_descriptor.hpp
+0
-1
include/ck/tensor_description/tensor_descriptor_helper.hpp
include/ck/tensor_description/tensor_descriptor_helper.hpp
+0
-1
include/ck/tensor_description/tensor_space_filling_curve.hpp
include/ck/tensor_description/tensor_space_filling_curve.hpp
+2
-5
No files found.
include/ck/ck.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
@@ -50,9 +49,9 @@
...
@@ -50,9 +49,9 @@
#define CK_BUFFER_RESOURCE_3RD_DWORD -1
#define CK_BUFFER_RESOURCE_3RD_DWORD -1
#elif defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) || defined(__gfx908__) || \
#elif defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) || defined(__gfx908__) || \
defined(__gfx90a__) || defined(__gfx940__) || defined(__gfx941__) || \
defined(__gfx90a__) || defined(__gfx940__) || defined(__gfx941__) || \
defined(__gfx942__) // for GPU code
defined(__gfx942__)
// for GPU code
#define CK_BUFFER_RESOURCE_3RD_DWORD 0x00020000
#define CK_BUFFER_RESOURCE_3RD_DWORD 0x00020000
#elif defined(__gfx1030__) // for GPU code
#elif defined(__gfx1030__)
// for GPU code
#define CK_BUFFER_RESOURCE_3RD_DWORD 0x31014000
#define CK_BUFFER_RESOURCE_3RD_DWORD 0x31014000
#elif defined(__gfx1100__) || defined(__gfx1101__) || defined(__gfx1102__) // for GPU code
#elif defined(__gfx1100__) || defined(__gfx1101__) || defined(__gfx1102__) // for GPU code
#define CK_BUFFER_RESOURCE_3RD_DWORD 0x31004000
#define CK_BUFFER_RESOURCE_3RD_DWORD 0x31004000
...
@@ -86,7 +85,7 @@
...
@@ -86,7 +85,7 @@
#endif
#endif
// WMMA instruction
// WMMA instruction
#ifndef __HIP_DEVICE_COMPILE__ // for host code
#ifndef __HIP_DEVICE_COMPILE__
// for host code
#define CK_USE_AMD_WMMA
#define CK_USE_AMD_WMMA
#elif defined(__gfx1100__) || defined(__gfx1101__) || defined(__gfx1102__) // for GPU code
#elif defined(__gfx1100__) || defined(__gfx1101__) || defined(__gfx1102__) // for GPU code
#define CK_USE_AMD_WMMA
#define CK_USE_AMD_WMMA
...
@@ -107,7 +106,7 @@
...
@@ -107,7 +106,7 @@
#elif defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx940__) || defined(__gfx941__) || \
#elif defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx940__) || defined(__gfx941__) || \
defined(__gfx942__) // for GPU code
defined(__gfx942__) // for GPU code
#define CK_USE_AMD_BUFFER_ATOMIC_ADD_FLOAT 1
#define CK_USE_AMD_BUFFER_ATOMIC_ADD_FLOAT 1
#else // for GPU code
#else
// for GPU code
#define CK_USE_AMD_BUFFER_ATOMIC_ADD_FLOAT 0
#define CK_USE_AMD_BUFFER_ATOMIC_ADD_FLOAT 0
#endif
#endif
...
...
include/ck/host_utility/device_prop.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/host_utility/hip_check_error.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/host_utility/io.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/host_utility/stream_utility.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/problem_transform/transform_forward_convolution3d_into_gemm_v4r4r4_ndhwc_kzyxc_ndhwk.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor/static_tensor.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor_description/cluster_descriptor.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor_description/multi_index_transform.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor_description/multi_index_transform_helper.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor_description/tensor_adaptor.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
@@ -109,13 +108,13 @@ struct TensorAdaptor
...
@@ -109,13 +108,13 @@ struct TensorAdaptor
__host__
__device__
static
constexpr
index_t
GetNumOfHiddenDimension
()
__host__
__device__
static
constexpr
index_t
GetNumOfHiddenDimension
()
{
{
constexpr
auto
all_low_dim_ids
=
unpack
(
constexpr
auto
all_low_dim_ids
=
[](
auto
&&
...
xs
)
constexpr
{
return
merge_sequences
(
xs
...);
},
unpack
(
[](
auto
&&
...
xs
)
constexpr
{
return
merge_sequences
(
xs
...);
},
LowerDimensionHiddenIdss
{});
LowerDimensionHiddenIdss
{});
constexpr
auto
all_up_dim_ids
=
unpack
(
constexpr
auto
all_up_dim_ids
=
[](
auto
&&
...
xs
)
constexpr
{
return
merge_sequences
(
xs
...);
},
unpack
(
[](
auto
&&
...
xs
)
constexpr
{
return
merge_sequences
(
xs
...);
},
UpperDimensionHiddenIdss
{});
UpperDimensionHiddenIdss
{});
constexpr
auto
all_dim_ids
=
merge_sequences
(
all_low_dim_ids
,
all_up_dim_ids
);
constexpr
auto
all_dim_ids
=
merge_sequences
(
all_low_dim_ids
,
all_up_dim_ids
);
...
@@ -339,8 +338,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
...
@@ -339,8 +338,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
TensorAdaptor1
::
GetLowerDimensionHiddenIdss
()[
itran
];
TensorAdaptor1
::
GetLowerDimensionHiddenIdss
()[
itran
];
// sequence in, sequence out
// sequence in, sequence out
constexpr
auto
low_dim_hidden_ids_1_mod
=
[
&
]()
constexpr
constexpr
auto
low_dim_hidden_ids_1_mod
=
[
&
]()
constexpr
{
{
auto
low_dim_hidden_ids_1_mod_
=
to_multi_index
(
low_dim_hidden_ids_1
);
auto
low_dim_hidden_ids_1_mod_
=
to_multi_index
(
low_dim_hidden_ids_1
);
// shift hidden id so every dim id is unique
// shift hidden id so every dim id is unique
...
@@ -362,8 +360,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
...
@@ -362,8 +360,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
});
});
return
low_dim_hidden_ids_1_mod_
;
return
low_dim_hidden_ids_1_mod_
;
}
}();
();
return
generate_sequence_v2
(
return
generate_sequence_v2
(
[
&
](
auto
i
)
constexpr
{
return
Number
<
low_dim_hidden_ids_1_mod
[
i
]
>
{};
},
[
&
](
auto
i
)
constexpr
{
return
Number
<
low_dim_hidden_ids_1_mod
[
i
]
>
{};
},
...
@@ -385,8 +382,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
...
@@ -385,8 +382,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
TensorAdaptor1
::
GetUpperDimensionHiddenIdss
()[
itran
];
TensorAdaptor1
::
GetUpperDimensionHiddenIdss
()[
itran
];
// sequence in, constexpr tuple out
// sequence in, constexpr tuple out
constexpr
auto
up_dim_hidden_ids_1_mod
=
[
&
]()
constexpr
constexpr
auto
up_dim_hidden_ids_1_mod
=
[
&
]()
constexpr
{
{
auto
up_dim_hidden_ids_1_mod_
=
to_multi_index
(
up_dim_hidden_ids_1
);
auto
up_dim_hidden_ids_1_mod_
=
to_multi_index
(
up_dim_hidden_ids_1
);
// shift hidden id
// shift hidden id
...
@@ -395,8 +391,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
...
@@ -395,8 +391,7 @@ __host__ __device__ constexpr auto chain_tensor_adaptors(const TensorAdaptor0& a
});
});
return
up_dim_hidden_ids_1_mod_
;
return
up_dim_hidden_ids_1_mod_
;
}
}();
();
// constexpr tuple to sequence
// constexpr tuple to sequence
return
generate_sequence_v2
(
return
generate_sequence_v2
(
...
...
include/ck/tensor_description/tensor_descriptor.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor_description/tensor_descriptor_helper.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
...
include/ck/tensor_description/tensor_space_filling_curve.hpp
View file @
dbe06911
// SPDX-License-Identifier: MIT
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
...
@@ -95,10 +94,8 @@ struct SpaceFillingCurve
...
@@ -95,10 +94,8 @@ struct SpaceFillingCurve
// Given tensor strides \p access_lengths, and 1D index of space-filling-curve, compute the
// Given tensor strides \p access_lengths, and 1D index of space-filling-curve, compute the
// idim-th element of multidimensional index.
// idim-th element of multidimensional index.
// All constexpr variables have to be captured by VALUE.
// All constexpr variables have to be captured by VALUE.
constexpr
auto
compute_index
=
[
idx_1d
,
access_strides
](
auto
idim
)
constexpr
constexpr
auto
compute_index
=
[
idx_1d
,
access_strides
](
auto
idim
)
constexpr
{
{
constexpr
auto
compute_index_impl
=
[
idx_1d
,
access_strides
](
auto
jdim
)
constexpr
{
constexpr
auto
compute_index_impl
=
[
idx_1d
,
access_strides
](
auto
jdim
)
constexpr
{
auto
res
=
idx_1d
.
value
;
auto
res
=
idx_1d
.
value
;
auto
id
=
0
;
auto
id
=
0
;
...
...
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