ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-19 22:39:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix ReLU formula (#61)

Browse Source 
* fix relu

* clean up

* clean up
This commit is contained in:
Chao Liu
2021-12-04 16:05:29 -06:00
committed by GitHub
parent 41cdd3801a
commit fd3d907a80
3 changed files with 156 additions and 120 deletions
Download Patch File Download Diff File Expand all files Collapse all files

201
example/1_gemm_xdl/gemm_xdl.cpp
View File

@@ -25,115 +25,76 @@ struct PassThrough
struct Relu
{
float alpha = 0.1;
// ReLU
template <typename T>
__host__ __device__ constexpr T operator()(T v) const
{
T tmp = alpha * v;
return tmp > 0 ? tmp : 0;
return v > 0 ? v : 0;
}
};
template <typename ADataType,
typename BDataType,
typename CDataType,
typename ALayout,
typename BLayout,
typename CLayout,
template <ck::index_t... Is>
using S = ck::Sequence<Is...>;
using ADataType = ck::half_t;
using BDataType = ck::half_t;
using CDataType = ck::half_t;
using AccDataType = float;
using ALayout = ck::tensor_layout::gemm::RowMajor;
using BLayout = ck::tensor_layout::gemm::ColumnMajor;
using CLayout = ck::tensor_layout::gemm::RowMajor;
using AOp = PassThrough;
using BOp = PassThrough;
using COp = Relu;
// Compilation parameters for NT problem
// clang-format off
using DeviceGemmInstance =
//#########################################| AData| BData| CData| AccData| ALayout| BLayout| CLayout| AElementwise| BElementwise| CElementwise| Block| MPer| NPer| K0Per| K1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| CThreadTransfer| CThreadTransfer| ABlockLds| BBlockLds|
//#########################################| Type| Type| Type| Type| | | | Operation| Operation| Operation| Size| Block| Block| Block| | XDL| XDL| Per| Per| ThreadSlice| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ThreadSlice| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| SrcDstVectorDim| DstScalar| AddExtraM| AddExtraN|
//#########################################| | | | | | | | | | | | | | | | | | Wave| Wave| Lengths_K0_N_K1| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| Lengths_K0_N_K1| Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerVector| | |
//#########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
ck::tensor_operation::device::DeviceGemmXdl< ADataType, BDataType, CDataType, AccDataType, ALayout, BLayout, CLayout, AOp, BOp, COp, 256, 256, 128, 4, 8, 32, 32, 4, 2, S<1, 4, 8>, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, S<1, 2, 8>, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 7, 1, true, true>;
// clang-format on
template <typename AType,
typename BType,
typename CType,
typename AElementwiseOperation,
typename BElementwiseOperation,
typename CElementwiseOperation>
struct DeviceGemmInstance;
template <typename AElementwiseOperation,
typename BElementwiseOperation,
typename CElementwiseOperation>
struct DeviceGemmInstance<ck::half_t,
ck::half_t,
ck::half_t,
ck::tensor_layout::gemm::RowMajor,
ck::tensor_layout::gemm::ColumnMajor,
ck::tensor_layout::gemm::RowMajor,
AElementwiseOperation,
BElementwiseOperation,
CElementwiseOperation>
static void host_verify(const Tensor<AType>& a_m_k,
const Tensor<BType>& b_k_n,
Tensor<CType>& c_m_n,
const AElementwiseOperation& a_element_op,
const BElementwiseOperation& b_element_op,
const CElementwiseOperation& c_element_op)
{
using F16 = ck::half_t;
using F32 = float;
auto f_mk_kn_mn = [&](auto m, auto n) {
const int K = a_m_k.mDesc.GetLengths()[1];
using Row = ck::tensor_layout::gemm::RowMajor;
using Col = ck::tensor_layout::gemm::ColumnMajor;
double v = 0;
template <ck::index_t... Is>
using S = ck::Sequence<Is...>;
for(int k = 0; k < K; ++k)
{
v += static_cast<const double>(a_element_op(a_m_k(m, k))) *
static_cast<const double>(b_element_op(b_k_n(k, n)));
}
using AOp = AElementwiseOperation;
using BOp = BElementwiseOperation;
using COp = CElementwiseOperation;
c_m_n(m, n) = c_element_op(v);
};
// Compilation parameters for NT problem
// clang-format off
using type =
//########################################| AData| BData| CData| AccData| ALayout| BLayout| CLayout| AElementwise| BElementwise| CElementwise| Block| MPer| NPer| K0Per| K1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| CThreadTransfer| CThreadTransfer| ABlockLds| BBlockLds|
//########################################| Type| Type| Type| Type| | | | Operation| Operation| Operation| Size| Block| Block| Block| | XDL| XDL| Per| Per| ThreadSlice| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ThreadSlice| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| SrcDstVectorDim| DstScalar| AddExtraM| AddExtraN|
//########################################| | | | | | | | | | | | | | | | | | Wave| Wave| Lengths_K0_N_K1| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| Lengths_K0_N_K1| Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerVector| | |
//########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
ck::tensor_operation::device::DeviceGemmXdl< F16, F16, F16, F32, Row, Col, Row, AOp, BOp, COp, 256, 256, 128, 4, 8, 32, 32, 4, 2, S<1, 4, 8>, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, S<1, 2, 8>, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 7, 1, true, true>;
// clang-format on
};
template <typename AElementwiseOperation,
typename BElementwiseOperation,
typename CElementwiseOperation>
struct DeviceGemmInstance<float,
float,
float,
ck::tensor_layout::gemm::RowMajor,
ck::tensor_layout::gemm::ColumnMajor,
ck::tensor_layout::gemm::RowMajor,
AElementwiseOperation,
BElementwiseOperation,
CElementwiseOperation>
{
using F16 = ck::half_t;
using F32 = float;
using Row = ck::tensor_layout::gemm::RowMajor;
using Col = ck::tensor_layout::gemm::ColumnMajor;
template <ck::index_t... Is>
using S = ck::Sequence<Is...>;
using AOp = AElementwiseOperation;
using BOp = BElementwiseOperation;
using COp = CElementwiseOperation;
// Compilation parameters for NT problem
// clang-format off
using type =
//########################################| AData| BData| CData| AccData| ALayout| BLayout| CLayout| AElementwise| BElementwise| CElementwise| Block| MPer| NPer| K0Per| K1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| CThreadTransfer| CThreadTransfer| ABlockLds| BBlockLds|
//########################################| Type| Type| Type| Type| | | | Operation| Operation| Operation| Size| Block| Block| Block| | XDL| XDL| Per| Per| ThreadSlice| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ThreadSlice| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| SrcDstVectorDim| DstScalar| AddExtraM| AddExtraN|
//########################################| | | | | | | | | | | | | | | | | | Wave| Wave| Lengths_K0_N_K1| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| Lengths_K0_N_K1| Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerVector| | |
//########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
ck::tensor_operation::device::DeviceGemmXdl< F32, F32, F32, F32, Row, Col, Row, AOp, BOp, COp, 256, 256, 128, 4, 4, 32, 32, 4, 2, S<1, 4, 4>, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, S<1, 2, 4>, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 7, 1, true, true>;
// clang-format on
};
make_ParallelTensorFunctor(f_mk_kn_mn,
c_m_n.mDesc.GetLengths()[0],
c_m_n.mDesc.GetLengths()[1])(std::thread::hardware_concurrency());
}
int main(int argc, char* argv[])
{
if(argc != 4)
{
printf("arg1: verification (0=no, 1=yes)\n");
printf("arg2: initialization (0=no init, 1=integer value, 2=decimal value)\n");
printf("arg3: run kernel # of times (>1)\n");
exit(0);
}
const bool do_verification = std::stoi(argv[1]);
const int init_method = std::stoi(argv[2]);
const int nrepeat = std::stoi(argv[3]);
bool do_verification = 0;
int init_method = 0;
int nrepeat = 5;
// GEMM shape
ck::index_t M = 3840;
@@ -144,15 +105,34 @@ int main(int argc, char* argv[])
ck::index_t StrideB = 4096;
ck::index_t StrideC = 4096;
// matrix data type
using ADataType = ck::half_t;
using BDataType = ck::half_t;
using CDataType = ck::half_t;
if(argc == 4)
{
M = std::stoi(argv[4]);
N = std::stoi(argv[5]);
K = std::stoi(argv[6]);
}
else if(argc == 10)
{
do_verification = std::stoi(argv[1]);
init_method = std::stoi(argv[2]);
nrepeat = std::stoi(argv[3]);
// matrix layout
using ALayout = ck::tensor_layout::gemm::RowMajor;
using BLayout = ck::tensor_layout::gemm::ColumnMajor;
using CLayout = ck::tensor_layout::gemm::RowMajor;
M = std::stoi(argv[4]);
N = std::stoi(argv[5]);
K = std::stoi(argv[6]);
StrideA = std::stoi(argv[7]);
StrideB = std::stoi(argv[8]);
StrideC = std::stoi(argv[9]);
}
else
{
printf("arg1: verification (0=no, 1=yes)\n");
printf("arg2: initialization (0=no init, 1=integer value, 2=decimal value)\n");
printf("arg3: run kernel # of times (>1)\n");
printf("arg4 to 9: M (256x), N(128x), K(32x), StrideA, StrideB, StrideC\n");
exit(0);
}
auto f_host_tensor_descriptor =
[](std::size_t row, std::size_t col, std::size_t stride, auto layout) {
@@ -198,16 +178,7 @@ int main(int argc, char* argv[])
c_m_n_device_buf.ToDevice(c_m_n_device_result.mData.data());
// do GEMM
auto gemm = typename DeviceGemmInstance<ADataType,
BDataType,
CDataType,
ALayout,
BLayout,
CLayout,
PassThrough,
PassThrough,
Relu>::type{};
auto gemm = DeviceGemmInstance{};
auto invoker = gemm.MakeInvoker();
auto argument = gemm.MakeArgument(static_cast<ADataType*>(a_m_k_device_buf.GetDeviceBuffer()),
static_cast<BDataType*>(b_k_n_device_buf.GetDeviceBuffer()),
@@ -218,9 +189,9 @@ int main(int argc, char* argv[])
StrideA,
StrideB,
StrideC,
PassThrough{},
PassThrough{},
Relu{});
AOp{},
BOp{},
COp{});
if(!gemm.IsSupportedArgument(argument))
{
@@ -233,7 +204,7 @@ int main(int argc, char* argv[])
std::size_t flop = std::size_t(2) * M * N * K;
std::size_t num_btype =
sizeof(ADataType) * M * K + sizeof(BDataType) * K * M + sizeof(CDataType) * M * N;
sizeof(ADataType) * M * K + sizeof(BDataType) * K * N + sizeof(CDataType) * M * N;
float tflops = static_cast<float>(flop) / 1.E9 / ave_time;
@@ -246,7 +217,7 @@ int main(int argc, char* argv[])
if(do_verification)
{
host_gemm_mk_kn_mn(a_m_k, b_k_n, c_m_n_host_result, PassThrough{}, PassThrough{}, Relu{});
host_verify(a_m_k, b_k_n, c_m_n_host_result, AOp{}, BOp{}, COp{});
check_error(c_m_n_host_result, c_m_n_device_result);
}

38
example/2_gemm_xdl_bias_relu_add/gemm_xdl_bias_relu_add.cpp
View File

@@ -20,10 +20,42 @@
// 0 in the "n" dimension
// assume C1 and C have same layout C
struct BiasReluAdd
{
template <typename T1, typename T2>
__host__ constexpr float operator()(float v0, T1 v1, T2 v2) const
{
float b = v0 + v1;
float c = b > 0 ? b : 0;
float d = c + v2;
return d;
}
template <typename T1, typename T2>
__device__ constexpr float operator()(float v0, T1 v1, T2 v2) const
{
#if 0
float a = v1 + v0;
float b = max(a, float(0));
float c = b + v2;
return c;
#else
float a = v1 + v2;
float b = v2;
float c = (v0 > -v1) ? a + v0 : v2;
return c;
#endif
}
};
// v0 is from A * B
// v1 is from C0
// v2 is from C1
struct BiasReluAdd
struct BiasLeakyReluAdd
{
template <typename T1, typename T2>
__host__ constexpr float operator()(float v0, T1 v1, T2 v2) const
@@ -51,7 +83,7 @@ struct BiasReluAdd
}
};
struct BiasRelu
struct BiasLeakyRelu
{
template <typename T1, typename T2>
__host__ constexpr float operator()(float v0, T1 v1, T2) const
@@ -99,7 +131,7 @@ struct BiasAdd
}
#elif 0
float alpha = 0.1;
float beta = 0.2;
float beta = 0.2;
float gamma = 0.3;
// wrong result

37
example/4_conv_xdl_bias_relu_add/conv_xdl_bias_relu_add.cpp
View File

@@ -23,7 +23,7 @@ struct PassThrough
}
};
struct BiasReluAdd
struct BiasLeakyReluAdd
{
template <typename T1, typename T2>
__host__ constexpr float operator()(float v0, T1 v1, T2 v2) const
@@ -97,7 +97,39 @@ struct BiasReluAdd
}
};
struct BiasRelu
struct BiasReluAdd
{
template <typename T1, typename T2>
__host__ constexpr float operator()(float v0, T1 v1, T2 v2) const
{
float b = v0 + v1;
float c = b > 0 ? b : 0;
float d = c + v2;
return d;
}
template <typename T1, typename T2>
__device__ constexpr float operator()(float v0, T1 v1, T2 v2) const
{
#if 0
float a = v1 + v0;
float b = max(a, float(0));
float c = b + v2;
return c;
#else
float a = v1 + v2;
float b = v2;
float c = (v0 > -v1) ? a + v0 : v2;
return c;
#endif
}
};
struct BiasLeakyRelu
{
template <typename T1, typename T2>
__host__ constexpr float operator()(float v0, T1 v1, T2) const
@@ -377,6 +409,7 @@ int main(int argc, char* argv[])
std::size_t num_btype = sizeof(InDataType) * (N * C * Hi * Wi) +
sizeof(WeiDataType) * (K * C * Y * X) +
sizeof(OutDataType) * (N * K * Ho * Wo) + sizeof(OutDataType) * (K) +
sizeof(OutDataType) * (N * K * Ho * Wo);
float tflops = static_cast<float>(flop) / 1.E9 / ave_time;