Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Refuce f16 #261

Merged
merged 19 commits into from
Aug 20, 2024
Merged

Refuce f16 #261

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
d5d09a6
Add f16 and bf16 instance files
CongMa13 Jul 15, 2024
de56fa9
f16 test
CongMa13 Aug 1, 2024
f2d5738
Reset input B of reduce tests
CongMa13 Aug 6, 2024
92e1cdf
Fixed bug in reduction_resource
CongMa13 Aug 8, 2024
fb0e5b6
fixed a bug of fillRand in reduction_resource
CongMa13 Aug 8, 2024
58d53a8
remove reduce all tests
CongMa13 Aug 8, 2024
ea55ca1
use f32 as compute type
CongMa13 Aug 8, 2024
1e0843f
Add another CK instance with InSrcVectorDim==1 which supports far rig…
CongMa13 Aug 9, 2024
8019d81
make the variable names readable
CongMa13 Aug 10, 2024
c6e62e7
Support reducing a tensor to a scalar
CongMa13 Aug 10, 2024
44debee
Rename label in reduction test config file
CongMa13 Aug 10, 2024
f6dd21f
Commnet out all test cases that beta != 0
CongMa13 Aug 12, 2024
a1b9782
Support independent tensor C and D
CongMa13 Aug 12, 2024
298d562
Support permutation of output of reduction
CongMa13 Aug 13, 2024
064f3c6
Add comment of permutation of output of reduction
CongMa13 Aug 14, 2024
0fdb159
comment out test cases that beta is not 0
CongMa13 Aug 16, 2024
ae562f2
Refactor reduction test
CongMa13 Aug 19, 2024
3e0d3c4
Refactor reduction
CongMa13 Aug 19, 2024
43d8de7
fixed a bug
CongMa13 Aug 19, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
42 changes: 42 additions & 0 deletions library/src/reduction/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,6 +33,48 @@ set(HIPTENSOR_REDUCTION_SOURCES
${CMAKE_CURRENT_SOURCE_DIR}/reduction_cpu_reference.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_cpu_reference_instances.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_1_1_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_2_1_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_2_2_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_3_1_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_3_2_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_3_3_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_1_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_2_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_3_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_4_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_1_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_2_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_3_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_4_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_5_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_1_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_2_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_3_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_4_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_5_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_6_f16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_1_1_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_2_1_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_2_2_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_3_1_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_3_2_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_3_3_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_1_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_2_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_3_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_4_4_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_1_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_2_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_3_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_4_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_5_5_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_1_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_2_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_3_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_4_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_5_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_6_6_bf16_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_1_1_f32_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_2_1_f32_f32_instance.cpp
${CMAKE_CURRENT_SOURCE_DIR}/reduction_solution_2_2_f32_f32_instance.cpp
Expand Down
9 changes: 8 additions & 1 deletion library/src/reduction/hiptensor_reduction.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -250,9 +250,16 @@ hiptensorStatus_t hiptensorReduction(const hiptensorHandle_t* handle,
auto ADataType = descA->mType;
auto DDataType = descD->mType;

auto internalTypeCompute = typeCompute;
if(typeCompute == HIPTENSOR_COMPUTE_16F || typeCompute == HIPTENSOR_COMPUTE_16BF)
{
// CK does not support f16 or bf16 as compute type
internalTypeCompute = HIPTENSOR_COMPUTE_32F;
}

// Query reduction solutions for the correct reduction operation and type
auto solutionQ = instances->querySolutions(ADataType,
typeCompute,
internalTypeCompute,
DDataType,
rankA,
numReduceDim,
Expand Down
13 changes: 10 additions & 3 deletions library/src/reduction/reduction_cpu_reference.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,15 +48,22 @@ hiptensorStatus_t hiptensorReductionReference(const void*
auto ADataType = descA->mType;
auto DDataType = descD->mType;

auto internalTypeCompute = typeCompute;
if(typeCompute == HIPTENSOR_COMPUTE_16F || typeCompute == HIPTENSOR_COMPUTE_16BF)
{
// CK does not support f16 or bf16 as compute type
internalTypeCompute = HIPTENSOR_COMPUTE_32F;
}

auto& instances = hiptensor::ReductionCpuReferenceInstances::instance();
auto solutionQ = instances->querySolutions(ADataType,
typeCompute,
internalTypeCompute,
DDataType,
rankA,
numReduceDim,
opReduce,
true, // @TODO hardcode
false); // @TODO hardcode
true, // propagateNan
false); // outputIndex

double alphaD;
if(alpha != nullptr)
Expand Down
235 changes: 161 additions & 74 deletions library/src/reduction/reduction_cpu_reference_instances.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,38 +28,38 @@
#include "reduction_cpu_reference_impl.hpp"
#include <hiptensor/internal/types.hpp>

#define REG_CPU_SOLUTION(dim_count, reduced_dim_count, type) \
registerSolutions(enumerateReferenceSolutions<type, \
type, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_ADD, \
true, \
false>()); \
registerSolutions(enumerateReferenceSolutions<type, \
type, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_MUL, \
true, \
false>()); \
registerSolutions(enumerateReferenceSolutions<type, \
type, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_MIN, \
true, \
false>()); \
registerSolutions(enumerateReferenceSolutions<type, \
type, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_MAX, \
true, \
#define REG_CPU_SOLUTION(dim_count, reduced_dim_count, type, computeType) \
cgmillette marked this conversation as resolved.
Show resolved Hide resolved
registerSolutions(enumerateReferenceSolutions<type, \
computeType, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_ADD, \
true, \
false>()); \
registerSolutions(enumerateReferenceSolutions<type, \
computeType, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_MUL, \
true, \
false>()); \
registerSolutions(enumerateReferenceSolutions<type, \
computeType, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_MIN, \
true, \
false>()); \
registerSolutions(enumerateReferenceSolutions<type, \
computeType, \
type, \
dim_count, \
reduced_dim_count, \
HIPTENSOR_OP_MAX, \
true, \
false>());

namespace hiptensor
Expand All @@ -83,49 +83,136 @@ namespace hiptensor
// HIPTENSOR_OP_ADD,
// true, // PropagateNan,
// false>()); // OutputIndex,
REG_CPU_SOLUTION(1, 1, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(2, 1, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(2, 2, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(3, 1, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(3, 2, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(3, 3, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(4, 1, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(4, 2, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(4, 3, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(4, 4, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(5, 1, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(5, 2, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(5, 3, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(5, 4, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(5, 5, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(6, 1, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(6, 2, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(6, 3, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(6, 4, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(6, 5, ck::half_t, ck::half_t);
REG_CPU_SOLUTION(6, 6, ck::half_t, ck::half_t);

REG_CPU_SOLUTION(1, 1, hiptensor::float32_t);
REG_CPU_SOLUTION(2, 1, hiptensor::float32_t);
REG_CPU_SOLUTION(2, 2, hiptensor::float32_t);
REG_CPU_SOLUTION(3, 1, hiptensor::float32_t);
REG_CPU_SOLUTION(3, 2, hiptensor::float32_t);
REG_CPU_SOLUTION(3, 3, hiptensor::float32_t);
REG_CPU_SOLUTION(4, 1, hiptensor::float32_t);
REG_CPU_SOLUTION(4, 2, hiptensor::float32_t);
REG_CPU_SOLUTION(4, 3, hiptensor::float32_t);
REG_CPU_SOLUTION(4, 4, hiptensor::float32_t);
REG_CPU_SOLUTION(5, 1, hiptensor::float32_t);
REG_CPU_SOLUTION(5, 2, hiptensor::float32_t);
REG_CPU_SOLUTION(5, 3, hiptensor::float32_t);
REG_CPU_SOLUTION(5, 4, hiptensor::float32_t);
REG_CPU_SOLUTION(5, 5, hiptensor::float32_t);
REG_CPU_SOLUTION(6, 1, hiptensor::float32_t);
REG_CPU_SOLUTION(6, 2, hiptensor::float32_t);
REG_CPU_SOLUTION(6, 3, hiptensor::float32_t);
REG_CPU_SOLUTION(6, 4, hiptensor::float32_t);
REG_CPU_SOLUTION(6, 5, hiptensor::float32_t);
REG_CPU_SOLUTION(6, 6, hiptensor::float32_t);
REG_CPU_SOLUTION(1, 1, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(2, 1, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(2, 2, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(3, 1, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(3, 2, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(3, 3, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(4, 1, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(4, 2, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(4, 3, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(4, 4, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(5, 1, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(5, 2, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(5, 3, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(5, 4, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(5, 5, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(6, 1, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(6, 2, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(6, 3, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(6, 4, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(6, 5, ck::bhalf_t, ck::bhalf_t);
REG_CPU_SOLUTION(6, 6, ck::bhalf_t, ck::bhalf_t);

REG_CPU_SOLUTION(1, 1, hiptensor::float64_t);
REG_CPU_SOLUTION(2, 1, hiptensor::float64_t);
REG_CPU_SOLUTION(2, 2, hiptensor::float64_t);
REG_CPU_SOLUTION(3, 1, hiptensor::float64_t);
REG_CPU_SOLUTION(3, 2, hiptensor::float64_t);
REG_CPU_SOLUTION(3, 3, hiptensor::float64_t);
REG_CPU_SOLUTION(4, 1, hiptensor::float64_t);
REG_CPU_SOLUTION(4, 2, hiptensor::float64_t);
REG_CPU_SOLUTION(4, 3, hiptensor::float64_t);
REG_CPU_SOLUTION(4, 4, hiptensor::float64_t);
REG_CPU_SOLUTION(5, 1, hiptensor::float64_t);
REG_CPU_SOLUTION(5, 2, hiptensor::float64_t);
REG_CPU_SOLUTION(5, 3, hiptensor::float64_t);
REG_CPU_SOLUTION(5, 4, hiptensor::float64_t);
REG_CPU_SOLUTION(5, 5, hiptensor::float64_t);
REG_CPU_SOLUTION(6, 1, hiptensor::float64_t);
REG_CPU_SOLUTION(6, 2, hiptensor::float64_t);
REG_CPU_SOLUTION(6, 3, hiptensor::float64_t);
REG_CPU_SOLUTION(6, 4, hiptensor::float64_t);
REG_CPU_SOLUTION(6, 5, hiptensor::float64_t);
REG_CPU_SOLUTION(6, 6, hiptensor::float64_t);
REG_CPU_SOLUTION(1, 1, ck::half_t, float);
REG_CPU_SOLUTION(2, 1, ck::half_t, float);
REG_CPU_SOLUTION(2, 2, ck::half_t, float);
REG_CPU_SOLUTION(3, 1, ck::half_t, float);
REG_CPU_SOLUTION(3, 2, ck::half_t, float);
REG_CPU_SOLUTION(3, 3, ck::half_t, float);
REG_CPU_SOLUTION(4, 1, ck::half_t, float);
REG_CPU_SOLUTION(4, 2, ck::half_t, float);
REG_CPU_SOLUTION(4, 3, ck::half_t, float);
REG_CPU_SOLUTION(4, 4, ck::half_t, float);
REG_CPU_SOLUTION(5, 1, ck::half_t, float);
REG_CPU_SOLUTION(5, 2, ck::half_t, float);
REG_CPU_SOLUTION(5, 3, ck::half_t, float);
REG_CPU_SOLUTION(5, 4, ck::half_t, float);
REG_CPU_SOLUTION(5, 5, ck::half_t, float);
REG_CPU_SOLUTION(6, 1, ck::half_t, float);
REG_CPU_SOLUTION(6, 2, ck::half_t, float);
REG_CPU_SOLUTION(6, 3, ck::half_t, float);
REG_CPU_SOLUTION(6, 4, ck::half_t, float);
REG_CPU_SOLUTION(6, 5, ck::half_t, float);
REG_CPU_SOLUTION(6, 6, ck::half_t, float);

REG_CPU_SOLUTION(1, 1, ck::bhalf_t, float);
REG_CPU_SOLUTION(2, 1, ck::bhalf_t, float);
REG_CPU_SOLUTION(2, 2, ck::bhalf_t, float);
REG_CPU_SOLUTION(3, 1, ck::bhalf_t, float);
REG_CPU_SOLUTION(3, 2, ck::bhalf_t, float);
REG_CPU_SOLUTION(3, 3, ck::bhalf_t, float);
REG_CPU_SOLUTION(4, 1, ck::bhalf_t, float);
REG_CPU_SOLUTION(4, 2, ck::bhalf_t, float);
REG_CPU_SOLUTION(4, 3, ck::bhalf_t, float);
REG_CPU_SOLUTION(4, 4, ck::bhalf_t, float);
REG_CPU_SOLUTION(5, 1, ck::bhalf_t, float);
REG_CPU_SOLUTION(5, 2, ck::bhalf_t, float);
REG_CPU_SOLUTION(5, 3, ck::bhalf_t, float);
REG_CPU_SOLUTION(5, 4, ck::bhalf_t, float);
REG_CPU_SOLUTION(5, 5, ck::bhalf_t, float);
REG_CPU_SOLUTION(6, 1, ck::bhalf_t, float);
REG_CPU_SOLUTION(6, 2, ck::bhalf_t, float);
REG_CPU_SOLUTION(6, 3, ck::bhalf_t, float);
REG_CPU_SOLUTION(6, 4, ck::bhalf_t, float);
REG_CPU_SOLUTION(6, 5, ck::bhalf_t, float);
REG_CPU_SOLUTION(6, 6, ck::bhalf_t, float);

REG_CPU_SOLUTION(1, 1, float, float);
REG_CPU_SOLUTION(2, 1, float, float);
REG_CPU_SOLUTION(2, 2, float, float);
REG_CPU_SOLUTION(3, 1, float, float);
REG_CPU_SOLUTION(3, 2, float, float);
REG_CPU_SOLUTION(3, 3, float, float);
REG_CPU_SOLUTION(4, 1, float, float);
REG_CPU_SOLUTION(4, 2, float, float);
REG_CPU_SOLUTION(4, 3, float, float);
REG_CPU_SOLUTION(4, 4, float, float);
REG_CPU_SOLUTION(5, 1, float, float);
REG_CPU_SOLUTION(5, 2, float, float);
REG_CPU_SOLUTION(5, 3, float, float);
REG_CPU_SOLUTION(5, 4, float, float);
REG_CPU_SOLUTION(5, 5, float, float);
REG_CPU_SOLUTION(6, 1, float, float);
REG_CPU_SOLUTION(6, 2, float, float);
REG_CPU_SOLUTION(6, 3, float, float);
REG_CPU_SOLUTION(6, 4, float, float);
REG_CPU_SOLUTION(6, 5, float, float);
REG_CPU_SOLUTION(6, 6, float, float);

REG_CPU_SOLUTION(1, 1, double, double);
REG_CPU_SOLUTION(2, 1, double, double);
REG_CPU_SOLUTION(2, 2, double, double);
REG_CPU_SOLUTION(3, 1, double, double);
REG_CPU_SOLUTION(3, 2, double, double);
REG_CPU_SOLUTION(3, 3, double, double);
REG_CPU_SOLUTION(4, 1, double, double);
REG_CPU_SOLUTION(4, 2, double, double);
REG_CPU_SOLUTION(4, 3, double, double);
REG_CPU_SOLUTION(4, 4, double, double);
REG_CPU_SOLUTION(5, 1, double, double);
REG_CPU_SOLUTION(5, 2, double, double);
REG_CPU_SOLUTION(5, 3, double, double);
REG_CPU_SOLUTION(5, 4, double, double);
REG_CPU_SOLUTION(5, 5, double, double);
REG_CPU_SOLUTION(6, 1, double, double);
REG_CPU_SOLUTION(6, 2, double, double);
REG_CPU_SOLUTION(6, 3, double, double);
REG_CPU_SOLUTION(6, 4, double, double);
REG_CPU_SOLUTION(6, 5, double, double);
REG_CPU_SOLUTION(6, 6, double, double);
}
} // namespace hiptensor
Loading
Loading