Skip to content

feat: add hardswish cpu/cuda kernel #73

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

Merged
Parent: Dev
merged 3 commits into from
Jan 18, 2024
Merged

feat: add hardswish cpu/cuda kernel #73

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
3998833
feat: add hardswish cpu/cuda kernel
bitzyz Jan 11, 2024
1ce6d81
fix: fix mod kernel
kilinchange Jan 11, 2024
afdfe93
style: 整理 hardswish 计算
YdrMaster Jan 18, 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
1 change: 1 addition & 0 deletions src/04kernel/include/kernel/collectors/simple_unary.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@ namespace refactor::kernel {
Erf,
Neg,
Not,
HardSwish,
};

std::string_view unaryName(SimpleUnaryType type);
Expand Down
1 change: 1 addition & 0 deletions src/04kernel/src/collectors/simple_unary.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,7 @@ namespace refactor::kernel {
CASE(Erf);
CASE(Neg);
CASE(Not);
CASE(HardSwish);
default:
UNREACHABLE();
}
Expand Down
8 changes: 4 additions & 4 deletions src/04kernel/src/kernels/simple_binary/cpu_kernel.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -137,11 +137,11 @@ namespace refactor::kernel {
switch (dataType.internal) {
CASE_DT(std::fmod(a, b), F32);
CASE_DT(a % b, U8);
CASE_DT(a % b < 0 ? (a % b + b) : (a % b), I8);
CASE_DT(static_cast<int8_t>(std::fmod(a, b)), I8);
CASE_DT(a % b, U16);
CASE_DT(a % b < 0 ? (a % b + b) : (a % b), I16);
CASE_DT(a % b < 0 ? (a % b + b) : (a % b), I32);
CASE_DT(a % b < 0 ? (a % b + b) : (a % b), I64);
CASE_DT(static_cast<int16_t>(std::fmod(a, b)), I16);
CASE_DT(static_cast<int32_t>(std::fmod(a, b)), I32);
CASE_DT(static_cast<int64_t>(std::fmod(a, b)), I64);
CASE_DT(std::fmod(a, b), F64);
CASE_DT(a % b, U32);
CASE_DT(a % b, U64);
Expand Down
12 changes: 8 additions & 4 deletions src/04kernel/src/kernels/simple_binary/cuda_kernel.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -158,14 +158,18 @@ extern "C" __global__ void kernel(
case SimpleBinaryType::Fmod:
switch (dt) {
case DataType::U8:
case DataType::I8:
case DataType::U16:
case DataType::U32:
case DataType::U64:
return "a % b";
case DataType::I8:
return "static_cast<char>(fmodf(a, b))";
case DataType::I16:
return "static_cast<short>(fmodf(a, b))";
case DataType::I32:
return "static_cast<int>(fmodf(a, b))";
case DataType::I64:
case DataType::U32:
case DataType::U64:
return "a % b < 0 ? (a % b + b) : (a % b)";
return "static_cast<long long>(fmodf(a, b))";
case DataType::F32:
return "fmodf(a, b)";
case DataType::FP16:
Expand Down
14 changes: 14 additions & 0 deletions src/04kernel/src/kernels/simple_unary/cpu_kernel.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@ namespace refactor::kernel {
Op::Tanh,
Op::Neg,
Op::Erf,
Op::HardSwish,
};
return supportedOp.contains(op) && a.dataType.isCpuNumberic()
? std::make_unique<K>(op, a.dataType, a.elementsSize())
Expand Down Expand Up @@ -49,6 +50,12 @@ namespace refactor::kernel {
using M = std::conditional_t<sizeof(T) <= 4, float, double>;
return static_cast<T>(std::tanh(static_cast<M>(x)));
}
template<class T> auto hardswishFun(T x) noexcept -> T {
auto mid = x / 6.f + .5f;
return (mid <= 0) ? 0
: (1 <= mid) ? x
: x * mid;
}
auto copyForUnsigned(size_t n) noexcept -> Routine {
return [n](runtime::Resources &, void *workspace, void const *const *inputs, void *const *outputs) {
std::memcpy(outputs[0], inputs[0], n);
Expand Down Expand Up @@ -171,6 +178,13 @@ namespace refactor::kernel {
default:
UNREACHABLE();
}
case Op::HardSwish:
switch (dataType) {
CASE(hardswishFun, F32);
CASE(hardswishFun, F64);
default:
UNREACHABLE();
}
default:
UNREACHABLE();
}
Expand Down
7 changes: 6 additions & 1 deletion src/04kernel/src/kernels/simple_unary/cuda_kernel.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ namespace refactor::kernel {
static const std::unordered_set<Op>
supportedOp{Op::Abs, Op::Relu, Op::Sqrt,
Op::Sigmoid, Op::Tanh, Op::Neg,
Op::Erf};
Op::Erf, Op::HardSwish};
#ifndef USE_CUDA
return nullptr;
#endif
Expand Down Expand Up @@ -154,6 +154,11 @@ extern "C" __global__ void kernel(
{__(Op::Erf, DT::I64 ), "erf(static_cast<double>(x))"},
{__(Op::Erf, DT::FP16), "__float2half(erff(__half2float(x)))"},
{__(Op::Erf, DT::BF16), "__float2bfloat16(erff(__bfloat162float(x)))"},

{__(Op::HardSwish, DT::F32 ), "x * fmaxf(0.f, fminf(1.f, fmaf(1.f/6.f, x, 0.5f)))"},
{__(Op::HardSwish, DT::FP16), "x * __hmax(CUDART_ZERO_FP16, __hmin(CUDART_ONE_FP16, hrcp(__float2half(6.f)) * x + hrcp(__float2half(2.f))))"},
{__(Op::HardSwish, DT::F64 ), "x * fmax(0.0, fmin(1.0, fma(1.0/6.0, x, 0.5)))"},

};
// clang-format on

Expand Down
2 changes: 1 addition & 1 deletion src/04kernel/test/kernels/simple_binary/test_binary_cpu.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,7 @@ TEST(kernel, BinaryCpu) {
testBinaryCPU(SimpleBinaryType::Mul, [](float a, float b) { return a * b; });
testBinaryCPU(SimpleBinaryType::Div, [](float a, float b) { return a / b; });
testModCPU(SimpleBinaryType::Mod, [](int a, int b) { return a % b; });
testFmodWithI32CPU(SimpleBinaryType::Fmod, [](int a, int b) { return a % b < 0 ? (a % b + b) : (a % b); });
testFmodWithI32CPU(SimpleBinaryType::Fmod, [](int a, int b) { return static_cast<int32_t>(std::fmod(a, b)); });
testBinaryCPU(SimpleBinaryType::Fmod, [](float a, float b) { return std::fmod(a, b); });
}

Expand Down
29 changes: 28 additions & 1 deletion src/04kernel/test/kernels/simple_unary/test_cpu.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,8 @@
using namespace refactor;
using namespace kernel;

using VecFloat = std::vector<float>;

static void testOp(SimpleUnaryType opType, float check(float)) {
// build routine
auto dataTensor = Tensor::share(DataType::F32, Shape{20, 30, 50});
Expand All @@ -12,7 +14,7 @@ static void testOp(SimpleUnaryType opType, float check(float)) {
auto res = runtime::Resources();
auto routine = kernel->lower(res).routine;
// put input data
std::vector<float> data(dataTensor->elementsSize());
VecFloat data(dataTensor->elementsSize());
for (auto i : range0_(data.size())) { data[i] = i * 1e-4f; }
auto result = data;
// inference
Expand All @@ -27,9 +29,34 @@ static void testOp(SimpleUnaryType opType, float check(float)) {
}
}

static void testOpWithData(SimpleUnaryType opType, const VecFloat &data) {
// build routine
auto dataTensor = Tensor::share(DataType::F32, Shape{2, 3});
auto kernel = SimpleUnaryCpu::build(opType, *dataTensor);
ASSERT_TRUE(kernel);
auto res = runtime::Resources();
auto routine = kernel->lower(res).routine;
// put input data
VecFloat inputdata(dataTensor->elementsSize());
for (auto i : range0_(inputdata.size())) { inputdata[i] = i; }
auto result = inputdata;
// inference
{
void const *inputs[]{result.data()};
void *outputs[]{result.data()};
routine(res, nullptr, inputs, outputs);
}
// check
for (auto i : range0_(inputdata.size())) {
EXPECT_NEAR(data[i], result[i], 1e-5);
}
}

TEST(kernel, SimpleUnaryCpu) {
testOp(SimpleUnaryType::Abs, std::abs);
testOp(SimpleUnaryType::Sqrt, std::sqrt);
testOp(SimpleUnaryType::Tanh, std::tanh);
testOp(SimpleUnaryType::Erf, std::erf);
testOpWithData(SimpleUnaryType::HardSwish,
VecFloat{0.000000, 0.666667, 1.666667, 3.000000, 4.000000, 5.000000});
}
1 change: 1 addition & 0 deletions src/04kernel/test/kernels/simple_unary/test_cuda.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,7 @@ TEST(kernel, SimpleUnaryCuda) {
testOp(SimpleUnaryType::Sigmoid);
testOp(SimpleUnaryType::Tanh);
testOp(SimpleUnaryType::Erf);
testOp(SimpleUnaryType::HardSwish);
}

#endif
6 changes: 6 additions & 0 deletions src/05computation/src/operators/simple_unary.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,6 +81,10 @@ namespace refactor::computation {
static uint8_t ID = 19;
return reinterpret_cast<size_t>(&ID);
}
case SimpleUnaryType::HardSwish: {
static uint8_t ID = 20;
return reinterpret_cast<size_t>(&ID);
}
default:
UNREACHABLE();
}
Expand Down Expand Up @@ -128,6 +132,8 @@ namespace refactor::computation {
return "Neg";
case SimpleUnaryType::Not:
return "Not";
case SimpleUnaryType::HardSwish:
return "HardSwish";
default:
UNREACHABLE();
}
Expand Down
2 changes: 2 additions & 0 deletions src/07onnx/src/operators.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -95,6 +95,7 @@ namespace refactor::onnx {
REGISTER(And , SimpleBinary );
REGISTER(Or , SimpleBinary );
REGISTER(Xor , SimpleBinary );
REGISTER(Mod , SimpleBinary );
REGISTER(Abs , SimpleUnary );
REGISTER(Acos , SimpleUnary );
REGISTER(Acosh , SimpleUnary );
Expand All @@ -116,6 +117,7 @@ namespace refactor::onnx {
REGISTER(Not , SimpleUnary );
REGISTER(Neg , SimpleUnary );
REGISTER(Identity , SimpleUnary );
REGISTER(HardSwish , SimpleUnary );
REGISTER(Slice , Slice );
REGISTER(Softmax , Softmax );
REGISTER(Split , Split );
Expand Down
9 changes: 8 additions & 1 deletion src/07onnx/src/operators/simple_unary.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,7 @@ namespace refactor::onnx {
opType == "onnx::Not" ? Ty::Not :
opType == "onnx::Neg" ? Ty::Neg :
opType == "onnx::Identity"? Ty::Identity:
opType == "onnx::HardSwish" ? Ty::HardSwish :
UNREACHABLEX(Ty, "Unsupported unary operator: {}", opType);
// clang-format on

Expand Down Expand Up @@ -129,6 +130,10 @@ namespace refactor::onnx {
static uint8_t ID = 21;
return reinterpret_cast<size_t>(&ID);
}
case Ty::HardSwish: {
static uint8_t ID = 22;
return reinterpret_cast<size_t>(&ID);
}
default:
UNREACHABLE();
}
Expand Down Expand Up @@ -159,6 +164,7 @@ namespace refactor::onnx {
case Ty::Not : return "onnx::Not";
case Ty::Neg : return "onnx::Neg";
case Ty::Identity : return "onnx::Identity";
case Ty::HardSwish : return "onnx::HardSwish";
default: UNREACHABLE();
}
// clang-format on
Expand Down Expand Up @@ -187,7 +193,7 @@ namespace refactor::onnx {
Ty::Atan, Ty::Atanh,
Ty::Cos, Ty::Cosh,
Ty::Sin, Ty::Sinh,
Ty::Tan},
Ty::Tan, Ty::HardSwish},
{Ty::Tanh, Ty::Sqrt, Ty::Sigmoid, Ty::Log},
{Ty::Neg},
{Ty::Identity}};
Expand Down Expand Up @@ -287,6 +293,7 @@ namespace refactor::onnx {
case Ty::Not : type_ = Ty_::Not ; break;
case Ty::Neg : type_ = Ty_::Neg ; break;
case Ty::Identity : return std::make_unique<computation::Identity>();
case Ty::HardSwish : type_ = Ty_::HardSwish ; break;
default: UNREACHABLE();
}
// clang-format on
Expand Down
17 changes: 9 additions & 8 deletions src/07onnx/src/operators/simple_unary.hh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,18 +16,19 @@ namespace refactor::onnx {
Atanh,
Cos,
Cosh,
Sin,
Sinh,
Tan,
Tanh,
Relu,
Sqrt,
Sigmoid,
Erf,
HardSwish,
Identity,
Log,
Not,
Neg,
Identity,
Relu,
Sin,
Sinh,
Sqrt,
Sigmoid,
Tan,
Tanh,
};

struct SimpleUnary final : public Operator {
Expand Down
14 changes: 14 additions & 0 deletions src/07onnx/test/test_simple_unary.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,4 +22,18 @@ TEST(infer, SimpleUnary) {
ASSERT_EQ(y->dataType, DataType::F32);
ASSERT_EQ(y->shape, (Shape{DimExpr(2), DimExpr(3)}));
}
{
// HardSwish Test
auto edges = Edges{
{Tensor::share(DataType::F32, Shape{DimExpr(2), DimExpr(3)}, {}), ""},
};
count_t inputs[]{0};
auto infered = SimpleUnary(SimpleUnaryType::HardSwish).infer(TensorRefs(edges, inputs), {true});
ASSERT_TRUE(infered.isOk());
auto outputs = std::move(infered.unwrap());
ASSERT_EQ(outputs.size(), 1);
auto y = std::move(outputs[0]);
ASSERT_EQ(y->dataType, DataType::F32);
ASSERT_EQ(y->shape, (Shape{DimExpr(2), DimExpr(3)}));
}
}