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
Commits
66fd0f8d
Commit
66fd0f8d
authored
Apr 19, 2022
by
Wen-Heng (Jack) Chung
Browse files
XXX only enable gemm.
parent
4221505d
Changes
5
Show whitespace changes
Inline
Side-by-side
Showing
5 changed files
with
151 additions
and
151 deletions
+151
-151
example/CMakeLists.txt
example/CMakeLists.txt
+15
-15
library/src/tensor_operation_instance/gpu/CMakeLists.txt
library/src/tensor_operation_instance/gpu/CMakeLists.txt
+17
-17
profiler/CMakeLists.txt
profiler/CMakeLists.txt
+29
-29
profiler/src/profiler.cpp
profiler/src/profiler.cpp
+78
-78
test/CMakeLists.txt
test/CMakeLists.txt
+12
-12
No files found.
example/CMakeLists.txt
View file @
66fd0f8d
...
@@ -27,18 +27,18 @@ function(add_example_executable EXAMPLE_NAME)
...
@@ -27,18 +27,18 @@ function(add_example_executable EXAMPLE_NAME)
endfunction
(
add_example_executable EXAMPLE_NAME
)
endfunction
(
add_example_executable EXAMPLE_NAME
)
add_subdirectory
(
01_gemm
)
add_subdirectory
(
01_gemm
)
add_subdirectory
(
02_gemm_alpha_beta
)
#
add_subdirectory(02_gemm_alpha_beta)
add_subdirectory
(
03_gemm_bias_relu
)
#
add_subdirectory(03_gemm_bias_relu)
add_subdirectory
(
04_gemm_bias_relu_add
)
#
add_subdirectory(04_gemm_bias_relu_add)
add_subdirectory
(
06_conv2d_fwd_bias_relu
)
#
add_subdirectory(06_conv2d_fwd_bias_relu)
add_subdirectory
(
07_conv2d_fwd_bias_relu_add
)
#
add_subdirectory(07_conv2d_fwd_bias_relu_add)
add_subdirectory
(
09_convnd_fwd
)
#
add_subdirectory(09_convnd_fwd)
add_subdirectory
(
10_conv2d_bwd_data
)
#
add_subdirectory(10_conv2d_bwd_data)
add_subdirectory
(
11_conv2d_bwd_weight
)
#
add_subdirectory(11_conv2d_bwd_weight)
add_subdirectory
(
12_reduce
)
#
add_subdirectory(12_reduce)
add_subdirectory
(
13_pool2d_fwd
)
#
add_subdirectory(13_pool2d_fwd)
add_subdirectory
(
14_gemm_xdl_requant_relu_requant
)
#
add_subdirectory(14_gemm_xdl_requant_relu_requant)
add_subdirectory
(
17_convnd_bwd_data_xdl
)
#
add_subdirectory(17_convnd_bwd_data_xdl)
add_subdirectory
(
15_grouped_gemm
)
#
add_subdirectory(15_grouped_gemm)
add_subdirectory
(
16_gemm_reduce
)
#
add_subdirectory(16_gemm_reduce)
add_subdirectory
(
18_batched_gemm_reduce
)
#
add_subdirectory(18_batched_gemm_reduce)
library/src/tensor_operation_instance/gpu/CMakeLists.txt
View file @
66fd0f8d
...
@@ -24,20 +24,20 @@ function(add_instance_library INSTANCE_NAME)
...
@@ -24,20 +24,20 @@ function(add_instance_library INSTANCE_NAME)
endfunction
(
add_instance_library INSTANCE_NAME
)
endfunction
(
add_instance_library INSTANCE_NAME
)
add_subdirectory
(
gemm
)
add_subdirectory
(
gemm
)
add_subdirectory
(
gemm_bias2d
)
#
add_subdirectory(gemm_bias2d)
add_subdirectory
(
gemm_bias_relu
)
#
add_subdirectory(gemm_bias_relu)
add_subdirectory
(
gemm_bias_relu_add
)
#
add_subdirectory(gemm_bias_relu_add)
add_subdirectory
(
gemm_reduce
)
#
add_subdirectory(gemm_reduce)
add_subdirectory
(
batched_gemm
)
#
add_subdirectory(batched_gemm)
add_subdirectory
(
conv1d_fwd
)
#
add_subdirectory(conv1d_fwd)
add_subdirectory
(
conv2d_fwd
)
#
add_subdirectory(conv2d_fwd)
add_subdirectory
(
conv3d_fwd
)
#
add_subdirectory(conv3d_fwd)
add_subdirectory
(
conv2d_fwd_bias_relu
)
#
add_subdirectory(conv2d_fwd_bias_relu)
add_subdirectory
(
conv2d_fwd_bias_relu_add
)
#
add_subdirectory(conv2d_fwd_bias_relu_add)
add_subdirectory
(
conv2d_fwd_bias_relu_atomic_add
)
#
add_subdirectory(conv2d_fwd_bias_relu_atomic_add)
add_subdirectory
(
conv2d_bwd_data
)
#
add_subdirectory(conv2d_bwd_data)
add_subdirectory
(
reduce
)
#
add_subdirectory(reduce)
add_subdirectory
(
convnd_bwd_data
)
#
add_subdirectory(convnd_bwd_data)
add_subdirectory
(
grouped_gemm
)
#
add_subdirectory(grouped_gemm)
add_subdirectory
(
conv2d_bwd_weight
)
#
add_subdirectory(conv2d_bwd_weight)
add_subdirectory
(
batched_gemm_reduce
)
#
add_subdirectory(batched_gemm_reduce)
profiler/CMakeLists.txt
View file @
66fd0f8d
...
@@ -24,38 +24,38 @@ include_directories(BEFORE
...
@@ -24,38 +24,38 @@ include_directories(BEFORE
set
(
PROFILER_SOURCE
set
(
PROFILER_SOURCE
src/profiler.cpp
src/profiler.cpp
src/profile_gemm.cpp
src/profile_gemm.cpp
src/profile_gemm_bias_2d.cpp
#
src/profile_gemm_bias_2d.cpp
src/profile_gemm_bias_relu.cpp
#
src/profile_gemm_bias_relu.cpp
src/profile_gemm_bias_relu_add.cpp
#
src/profile_gemm_bias_relu_add.cpp
src/profile_gemm_reduce.cpp
#
src/profile_gemm_reduce.cpp
src/profile_batched_gemm.cpp
#
src/profile_batched_gemm.cpp
src/profile_conv_fwd.cpp
#
src/profile_conv_fwd.cpp
src/profile_conv_fwd_bias_relu.cpp
#
src/profile_conv_fwd_bias_relu.cpp
src/profile_conv_fwd_bias_relu_add.cpp
#
src/profile_conv_fwd_bias_relu_add.cpp
src/profile_conv_fwd_bias_relu_atomic_add.cpp
#
src/profile_conv_fwd_bias_relu_atomic_add.cpp
src/profile_convnd_bwd_data.cpp
#
src/profile_convnd_bwd_data.cpp
src/profile_reduce.cpp
#
src/profile_reduce.cpp
src/profile_grouped_gemm.cpp
#
src/profile_grouped_gemm.cpp
src/profile_conv_bwd_weight.cpp
#
src/profile_conv_bwd_weight.cpp
src/profile_batched_gemm_reduce.cpp
#
src/profile_batched_gemm_reduce.cpp
)
)
add_executable
(
ckProfiler
${
PROFILER_SOURCE
}
)
add_executable
(
ckProfiler
${
PROFILER_SOURCE
}
)
target_link_libraries
(
ckProfiler PRIVATE host_tensor
)
target_link_libraries
(
ckProfiler PRIVATE host_tensor
)
target_link_libraries
(
ckProfiler PRIVATE device_gemm_reduce_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_gemm_reduce_instance)
target_link_libraries
(
ckProfiler PRIVATE device_gemm_instance
)
target_link_libraries
(
ckProfiler PRIVATE device_gemm_instance
)
target_link_libraries
(
ckProfiler PRIVATE device_gemm_bias2d_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_gemm_bias2d_instance)
target_link_libraries
(
ckProfiler PRIVATE device_gemm_bias_relu_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_gemm_bias_relu_instance)
target_link_libraries
(
ckProfiler PRIVATE device_gemm_bias_relu_add_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_gemm_bias_relu_add_instance)
target_link_libraries
(
ckProfiler PRIVATE device_batched_gemm_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_batched_gemm_instance)
target_link_libraries
(
ckProfiler PRIVATE device_conv2d_fwd_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_instance)
target_link_libraries
(
ckProfiler PRIVATE device_conv2d_fwd_bias_relu_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_bias_relu_instance)
target_link_libraries
(
ckProfiler PRIVATE device_conv2d_fwd_bias_relu_add_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_bias_relu_add_instance)
target_link_libraries
(
ckProfiler PRIVATE device_conv2d_fwd_bias_relu_atomic_add_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_conv2d_fwd_bias_relu_atomic_add_instance)
target_link_libraries
(
ckProfiler PRIVATE device_convnd_bwd_data_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_convnd_bwd_data_instance)
target_link_libraries
(
ckProfiler PRIVATE device_reduce_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_reduce_instance)
target_link_libraries
(
ckProfiler PRIVATE device_reduce_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_reduce_instance)
target_link_libraries
(
ckProfiler PRIVATE device_grouped_gemm_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_grouped_gemm_instance)
target_link_libraries
(
ckProfiler PRIVATE device_conv2d_bwd_weight_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_conv2d_bwd_weight_instance)
target_link_libraries
(
ckProfiler PRIVATE device_batched_gemm_reduce_instance
)
#
target_link_libraries(ckProfiler PRIVATE device_batched_gemm_reduce_instance)
profiler/src/profiler.cpp
View file @
66fd0f8d
...
@@ -5,20 +5,20 @@
...
@@ -5,20 +5,20 @@
#include <cstring>
#include <cstring>
int
profile_gemm
(
int
,
char
*
[]);
int
profile_gemm
(
int
,
char
*
[]);
int
profile_gemm_bias_2d
(
int
,
char
*
[]);
//
int profile_gemm_bias_2d(int, char*[]);
int
profile_gemm_bias_relu
(
int
,
char
*
[]);
//
int profile_gemm_bias_relu(int, char*[]);
int
profile_gemm_bias_relu_add
(
int
,
char
*
[]);
//
int profile_gemm_bias_relu_add(int, char*[]);
int
profile_gemm_reduce
(
int
,
char
*
[]);
//
int profile_gemm_reduce(int, char*[]);
int
profile_batched_gemm
(
int
,
char
*
[]);
//
int profile_batched_gemm(int, char*[]);
int
profile_grouped_gemm
(
int
,
char
*
[]);
//
int profile_grouped_gemm(int, char*[]);
int
profile_conv_fwd
(
int
,
char
*
[]);
//
int profile_conv_fwd(int, char*[]);
int
profile_conv_fwd_bias_relu
(
int
,
char
*
[]);
//
int profile_conv_fwd_bias_relu(int, char*[]);
int
profile_conv_fwd_bias_relu_add
(
int
,
char
*
[]);
//
int profile_conv_fwd_bias_relu_add(int, char*[]);
int
profile_conv_fwd_bias_relu_atomic_add
(
int
,
char
*
[]);
//
int profile_conv_fwd_bias_relu_atomic_add(int, char*[]);
int
profile_convnd_bwd_data
(
int
,
char
*
[],
int
);
//
int profile_convnd_bwd_data(int, char*[], int);
int
profile_reduce
(
int
,
char
*
[]);
//
int profile_reduce(int, char*[]);
int
profile_conv_bwd_weight
(
int
,
char
*
[]);
//
int profile_conv_bwd_weight(int, char*[]);
int
profile_batched_gemm_reduce
(
int
,
char
*
[]);
//
int profile_batched_gemm_reduce(int, char*[]);
int
main
(
int
argc
,
char
*
argv
[])
int
main
(
int
argc
,
char
*
argv
[])
{
{
...
@@ -26,70 +26,70 @@ int main(int argc, char* argv[])
...
@@ -26,70 +26,70 @@ int main(int argc, char* argv[])
{
{
return
profile_gemm
(
argc
,
argv
);
return
profile_gemm
(
argc
,
argv
);
}
}
else
if
(
strcmp
(
argv
[
1
],
"gemm_bias_2d"
)
==
0
)
//
else if(strcmp(argv[1], "gemm_bias_2d") == 0)
{
//
{
return
profile_gemm_bias_2d
(
argc
,
argv
);
//
return profile_gemm_bias_2d(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"gemm_bias_relu"
)
==
0
)
//
else if(strcmp(argv[1], "gemm_bias_relu") == 0)
{
//
{
return
profile_gemm_bias_relu
(
argc
,
argv
);
//
return profile_gemm_bias_relu(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"gemm_bias_relu_add"
)
==
0
)
//
else if(strcmp(argv[1], "gemm_bias_relu_add") == 0)
{
//
{
return
profile_gemm_bias_relu_add
(
argc
,
argv
);
//
return profile_gemm_bias_relu_add(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"gemm_reduce"
)
==
0
)
//
else if(strcmp(argv[1], "gemm_reduce") == 0)
{
//
{
return
profile_gemm_reduce
(
argc
,
argv
);
//
return profile_gemm_reduce(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"batched_gemm"
)
==
0
)
//
else if(strcmp(argv[1], "batched_gemm") == 0)
{
//
{
return
profile_batched_gemm
(
argc
,
argv
);
//
return profile_batched_gemm(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"batched_gemm_reduce"
)
==
0
)
//
else if(strcmp(argv[1], "batched_gemm_reduce") == 0)
{
//
{
return
profile_batched_gemm_reduce
(
argc
,
argv
);
//
return profile_batched_gemm_reduce(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"grouped_gemm"
)
==
0
)
//
else if(strcmp(argv[1], "grouped_gemm") == 0)
{
//
{
profile_grouped_gemm
(
argc
,
argv
);
//
profile_grouped_gemm(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv_fwd"
)
==
0
)
//
else if(strcmp(argv[1], "conv_fwd") == 0)
{
//
{
return
profile_conv_fwd
(
argc
,
argv
);
//
return profile_conv_fwd(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv_fwd_bias_relu"
)
==
0
)
//
else if(strcmp(argv[1], "conv_fwd_bias_relu") == 0)
{
//
{
return
profile_conv_fwd_bias_relu
(
argc
,
argv
);
//
return profile_conv_fwd_bias_relu(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv_fwd_bias_relu_add"
)
==
0
)
//
else if(strcmp(argv[1], "conv_fwd_bias_relu_add") == 0)
{
//
{
return
profile_conv_fwd_bias_relu_add
(
argc
,
argv
);
//
return profile_conv_fwd_bias_relu_add(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv_fwd_bias_relu_atomic_add"
)
==
0
)
//
else if(strcmp(argv[1], "conv_fwd_bias_relu_atomic_add") == 0)
{
//
{
return
profile_conv_fwd_bias_relu_atomic_add
(
argc
,
argv
);
//
return profile_conv_fwd_bias_relu_atomic_add(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv1d_bwd_data"
)
==
0
)
//
else if(strcmp(argv[1], "conv1d_bwd_data") == 0)
{
//
{
return
profile_convnd_bwd_data
(
argc
,
argv
,
1
);
//
return profile_convnd_bwd_data(argc, argv, 1);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv2d_bwd_data"
)
==
0
)
//
else if(strcmp(argv[1], "conv2d_bwd_data") == 0)
{
//
{
return
profile_convnd_bwd_data
(
argc
,
argv
,
2
);
//
return profile_convnd_bwd_data(argc, argv, 2);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv3d_bwd_data"
)
==
0
)
//
else if(strcmp(argv[1], "conv3d_bwd_data") == 0)
{
//
{
return
profile_convnd_bwd_data
(
argc
,
argv
,
3
);
//
return profile_convnd_bwd_data(argc, argv, 3);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"reduce"
)
==
0
)
//
else if(strcmp(argv[1], "reduce") == 0)
{
//
{
return
profile_reduce
(
argc
,
argv
);
//
return profile_reduce(argc, argv);
}
//
}
else
if
(
strcmp
(
argv
[
1
],
"conv2d_bwd_weight"
)
==
0
)
//
else if(strcmp(argv[1], "conv2d_bwd_weight") == 0)
{
//
{
return
profile_conv_bwd_weight
(
argc
,
argv
);
//
return profile_conv_bwd_weight(argc, argv);
}
//
}
else
else
{
{
// clang-format off
// clang-format off
...
...
test/CMakeLists.txt
View file @
66fd0f8d
...
@@ -32,16 +32,16 @@ function(add_test_executable TEST_NAME)
...
@@ -32,16 +32,16 @@ function(add_test_executable TEST_NAME)
add_dependencies
(
check
${
TEST_NAME
}
)
add_dependencies
(
check
${
TEST_NAME
}
)
endfunction
(
add_test_executable TEST_NAME
)
endfunction
(
add_test_executable TEST_NAME
)
add_subdirectory
(
magic_number_division
)
#
add_subdirectory(magic_number_division)
add_subdirectory
(
space_filling_curve
)
#
add_subdirectory(space_filling_curve)
add_subdirectory
(
conv_util
)
#
add_subdirectory(conv_util)
add_subdirectory
(
reference_conv_fwd
)
#
add_subdirectory(reference_conv_fwd)
add_subdirectory
(
gemm
)
add_subdirectory
(
gemm
)
add_subdirectory
(
gemm_split_k
)
#
add_subdirectory(gemm_split_k)
add_subdirectory
(
gemm_reduce
)
#
add_subdirectory(gemm_reduce)
add_subdirectory
(
batched_gemm
)
#
add_subdirectory(batched_gemm)
add_subdirectory
(
batched_gemm_reduce
)
#
add_subdirectory(batched_gemm_reduce)
add_subdirectory
(
grouped_gemm
)
#
add_subdirectory(grouped_gemm)
add_subdirectory
(
convnd_fwd
)
#
add_subdirectory(convnd_fwd)
add_subdirectory
(
reduce
)
#
add_subdirectory(reduce)
add_subdirectory
(
conv2d_bwd_weight
)
#
add_subdirectory(conv2d_bwd_weight)
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