Skip to content

GitLab

"git@developer.sourcefind.cn:OpenDAS/fast_rnnt.git" did not exist on "912b5f554fa9c933476d4a071d1fe213c4e11616"
Unverified Commit 35e49f2d authored by zjing14's avatar zjing14 Committed by GitHub
Browse files

add g; fixed strides (#355)

parent de60d290
Hide whitespace changes
Inline Side-by-side
Showing with 279 additions and 250 deletions
example/25_gemm_bias_e_permute/CMakeLists.txt
View file @ 35e49f2d
add_example_executable(example_gemm_bias_e_permute_m3n2_xdl_fp16 gemm_bias_e_permute_m3n2_xdl_fp16.cpp) add_example_executable(example_gemm_bias_e_permute_g1m3n2k1_xdl_fp16 gemm_bias_e_permute_g1m3n2k1_xdl_fp16.cpp)
add_example_executable(example_gemm_bias_e_permute_m2n3_xdl_fp16 gemm_bias_e_permute_m2n3_xdl_fp16.cpp) add_example_executable(example_gemm_bias_e_permute_g1m2n3k1_xdl_fp16 gemm_bias_e_permute_g1m2n3k1_xdl_fp16.cpp)
example/25_gemm_bias_e_permute/gemm_bias_e_permute_m2n3_xdl_fp16.cpp → example/25_gemm_bias_e_permute/gemm_bias_e_permute_g1m2n3k1_xdl_fp16.cpp
View file @ 35e49f2d
...@@ -16,6 +16,8 @@ ...@@ -16,6 +16,8 @@
#include "ck/library/utility/host_tensor.hpp" #include "ck/library/utility/host_tensor.hpp"
#include "ck/library/utility/host_tensor_generator.hpp" #include "ck/library/utility/host_tensor_generator.hpp"
#include "ck/library/reference_tensor_operation/cpu/reference_gemm.hpp"
template <ck::index_t... Is> template <ck::index_t... Is>
using S = ck::Sequence<Is...>; using S = ck::Sequence<Is...>;
...@@ -33,7 +35,7 @@ using DDataType = F16; ...@@ -33,7 +35,7 @@ using DDataType = F16;
using DsDataType = ck::Tuple<DDataType>; using DsDataType = ck::Tuple<DDataType>;
using EDataType = F16; using EDataType = F16;
static constexpr ck::index_t NumDimG = 0; static constexpr ck::index_t NumDimG = 1;
static constexpr ck::index_t NumDimM = 2; static constexpr ck::index_t NumDimM = 2;
static constexpr ck::index_t NumDimN = 3; static constexpr ck::index_t NumDimN = 3;
static constexpr ck::index_t NumDimK = 1; static constexpr ck::index_t NumDimK = 1;
...@@ -69,30 +71,31 @@ template <ck::index_t NumDimM, ...@@ -69,30 +71,31 @@ template <ck::index_t NumDimM,
typename AElementwiseOperation, typename AElementwiseOperation,
typename BElementwiseOperation, typename BElementwiseOperation,
typename CDEElementwiseOperation, typename CDEElementwiseOperation,
ck::enable_if_t<NumDimM == 2 && NumDimN == 3 && NumDimK == 1, bool> = false> ck::enable_if_t<NumDimG == 1 && NumDimM == 2 && NumDimN == 3 && NumDimK == 1, bool> =
struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::BaseOperator false>
struct ReferenceContraction_G1_M2_N3_K1 : public ck::tensor_operation::device::BaseOperator
{ {
// Argument // Argument
struct Argument : public ck::tensor_operation::device::BaseArgument struct Argument : public ck::tensor_operation::device::BaseArgument
{ {
Argument(const Tensor<ADataType>& a_ms_ks, Argument(const Tensor<ADataType>& a_gs_ms_ks,
const Tensor<BDataType>& b_ns_ks, const Tensor<BDataType>& b_gs_ns_ks,
Tensor<EDataType>& e_ms_ns, Tensor<EDataType>& e_gs_ms_ns,
AElementwiseOperation a_element_op, AElementwiseOperation a_element_op,
BElementwiseOperation b_element_op, BElementwiseOperation b_element_op,
CDEElementwiseOperation cde_element_op) CDEElementwiseOperation cde_element_op)
: a_ms_ks_{a_ms_ks}, : a_gs_ms_ks_{a_gs_ms_ks},
b_ns_ks_{b_ns_ks}, b_gs_ns_ks_{b_gs_ns_ks},
e_ms_ns_{e_ms_ns}, e_gs_ms_ns_{e_gs_ms_ns},
a_element_op_{a_element_op}, a_element_op_{a_element_op},
b_element_op_{b_element_op}, b_element_op_{b_element_op},
cde_element_op_{cde_element_op} cde_element_op_{cde_element_op}
{ {
} }
const Tensor<ADataType>& a_ms_ks_; const Tensor<ADataType>& a_gs_ms_ks_;
const Tensor<BDataType>& b_ns_ks_; const Tensor<BDataType>& b_gs_ns_ks_;
Tensor<EDataType>& e_ms_ns_; Tensor<EDataType>& e_gs_ms_ns_;
AElementwiseOperation a_element_op_; AElementwiseOperation a_element_op_;
BElementwiseOperation b_element_op_; BElementwiseOperation b_element_op_;
...@@ -102,12 +105,12 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base ...@@ -102,12 +105,12 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base
// Invoker // Invoker
struct Invoker : public ck::tensor_operation::device::BaseInvoker struct Invoker : public ck::tensor_operation::device::BaseInvoker
{ {
using Argument = ReferenceContraction_M2_N3_K1::Argument; using Argument = ReferenceContraction_G1_M2_N3_K1::Argument;
float Run(const Argument& arg) float Run(const Argument& arg)
{ {
auto f_ms_ns = [&](auto m0, auto m1, auto n0, auto n1, auto n2) { auto f_gs_ms_ns = [&](auto g0, auto m0, auto m1, auto n0, auto n1, auto n2) {
const int K0 = arg.a_ms_ks_.mDesc.GetLengths()[2]; const int K0 = arg.a_gs_ms_ks_.mDesc.GetLengths()[3];
AccDataType v_acc = 0; AccDataType v_acc = 0;
...@@ -117,9 +120,10 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base ...@@ -117,9 +120,10 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base
AccDataType v_b; AccDataType v_b;
arg.a_element_op_( arg.a_element_op_(
v_a, ck::type_convert<const AccDataType>(arg.a_ms_ks_(m0, m1, k0))); v_a, ck::type_convert<const AccDataType>(arg.a_gs_ms_ks_(g0, m0, m1, k0)));
arg.b_element_op_( arg.b_element_op_(
v_b, ck::type_convert<const AccDataType>(arg.b_ns_ks_(n0, n1, n2, k0))); v_b,
ck::type_convert<const AccDataType>(arg.b_gs_ns_ks_(g0, n0, n1, n2, k0)));
v_acc += v_a * v_b; v_acc += v_a * v_b;
} }
...@@ -128,15 +132,16 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base ...@@ -128,15 +132,16 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base
arg.cde_element_op_(v_c, v_acc); arg.cde_element_op_(v_c, v_acc);
arg.e_ms_ns_(m0, m1, n0, n1, n2) = v_c; arg.e_gs_ms_ns_(g0, m0, m1, n0, n1, n2) = v_c;
}; };
make_ParallelTensorFunctor(f_ms_ns, make_ParallelTensorFunctor(f_gs_ms_ns,
arg.e_ms_ns_.mDesc.GetLengths()[0], arg.e_gs_ms_ns_.mDesc.GetLengths()[0],
arg.e_ms_ns_.mDesc.GetLengths()[1], arg.e_gs_ms_ns_.mDesc.GetLengths()[1],
arg.e_ms_ns_.mDesc.GetLengths()[2], arg.e_gs_ms_ns_.mDesc.GetLengths()[2],
arg.e_ms_ns_.mDesc.GetLengths()[3], arg.e_gs_ms_ns_.mDesc.GetLengths()[3],
arg.e_ms_ns_.mDesc.GetLengths()[4])( arg.e_gs_ms_ns_.mDesc.GetLengths()[4],
arg.e_gs_ms_ns_.mDesc.GetLengths()[5])(
std::thread::hardware_concurrency()); std::thread::hardware_concurrency());
return 0; return 0;
...@@ -160,14 +165,15 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base ...@@ -160,14 +165,15 @@ struct ReferenceContraction_M2_N3_K1 : public ck::tensor_operation::device::Base
return true; return true;
} }
static auto MakeArgument(const Tensor<ADataType>& a_ms_ks, static auto MakeArgument(const Tensor<ADataType>& a_gs_ms_ks,
const Tensor<BDataType>& b_ns_ks, const Tensor<BDataType>& b_gs_ns_ks,
Tensor<EDataType>& e_ms_ns, Tensor<EDataType>& e_gs_ms_ns,
AElementwiseOperation a_element_op, AElementwiseOperation a_element_op,
BElementwiseOperation b_element_op, BElementwiseOperation b_element_op,
CDEElementwiseOperation cde_element_op) CDEElementwiseOperation cde_element_op)
{ {
return Argument{a_ms_ks, b_ns_ks, e_ms_ns, a_element_op, b_element_op, cde_element_op}; return Argument{
a_gs_ms_ks, b_gs_ns_ks, e_gs_ms_ns, a_element_op, b_element_op, cde_element_op};
} }
static auto MakeInvoker() { return Invoker{}; } static auto MakeInvoker() { return Invoker{}; }
...@@ -196,28 +202,31 @@ int main(int argc, char* argv[]) ...@@ -196,28 +202,31 @@ int main(int argc, char* argv[])
int init_method = 1; int init_method = 1;
bool time_kernel = false; bool time_kernel = false;
ck::index_t G0 = 1;
ck::index_t M0 = 4; ck::index_t M0 = 4;
ck::index_t M1 = 256; ck::index_t M1 = 256;
ck::index_t N0 = 4; ck::index_t N0 = 4;
ck::index_t N1 = 8; ck::index_t N1 = 16;
ck::index_t N2 = 128; ck::index_t N2 = 32;
ck::index_t K0 = 256; ck::index_t K0 = 256;
// A[M0, M1, M2, K0] // A[M0, M1, M2, K0]
std::vector<ck::index_t> a_ms_ks_lengths{M0, M1, K0}; std::vector<ck::index_t> a_gs_ms_ks_lengths{G0, M0, M1, K0};
std::vector<ck::index_t> a_ms_ks_strides{M1 * K0, K0, 1}; std::vector<ck::index_t> a_gs_ms_ks_strides{M0 * M1 * K0, M1 * K0, K0, 1};
// B[N0, N1, K0] // B[N0, N1, K0]
std::vector<ck::index_t> b_ns_ks_lengths{N0, N1, N2, K0}; std::vector<ck::index_t> b_gs_ns_ks_lengths{G0, N0, N1, N2, K0};
std::vector<ck::index_t> b_ns_ks_strides{N1 * N2 * K0, N2 * K0, K0, 1}; std::vector<ck::index_t> b_gs_ns_ks_strides{N0 * N1 * N2 * K0, N1 * N2 * K0, N2 * K0, K0, 1};
// D[N0, M0, N1, M1, N2] // D[N0, M0, N1, M1, N2]
std::vector<ck::index_t> d_ms_ns_lengths{M0, M1, N0, N1, N2}; std::vector<ck::index_t> d_gs_ms_ns_lengths{G0, M0, M1, N0, N1, N2};
std::vector<ck::index_t> d_ms_ns_strides{0, 0, N1 * N2, N1, 1}; std::vector<ck::index_t> d_gs_ms_ns_strides{N0 * N1 * N2, 0, 0, N1 * N2, N2, 1};
// E[N0, M0, N1, M1, N2] // E[N0, M0, N1, M1, N2]
std::vector<ck::index_t> e_ms_ns_lengths{M0, M1, N0, N1, N2}; std::vector<ck::index_t> e_gs_ms_ns_lengths{G0, M0, M1, N0, N1, N2};
std::vector<ck::index_t> e_ms_ns_strides{N1 * M1 * N2, N2, M0 * N1 * M1 * N2, M1 * N2, 1}; std::vector<ck::index_t> e_gs_ms_ns_strides{
M0 * M1 * N0 * N1 * N2, N1 * M1 * N2, N2, M0 * N1 * M1 * N2, M1 * N2, 1};
if(argc == 1) if(argc == 1)
{ {
...@@ -237,50 +246,51 @@ int main(int argc, char* argv[]) ...@@ -237,50 +246,51 @@ int main(int argc, char* argv[])
exit(0); exit(0);
} }
Tensor<ADataType> a_ms_ks( Tensor<ADataType> a_gs_ms_ks(
std::vector<std::size_t>(a_ms_ks_lengths.begin(), a_ms_ks_lengths.end()), std::vector<std::size_t>(a_gs_ms_ks_lengths.begin(), a_gs_ms_ks_lengths.end()),
std::vector<std::size_t>(a_ms_ks_strides.begin(), a_ms_ks_strides.end())); std::vector<std::size_t>(a_gs_ms_ks_strides.begin(), a_gs_ms_ks_strides.end()));
Tensor<BDataType> b_ns_ks( Tensor<BDataType> b_gs_ns_ks(
std::vector<std::size_t>(b_ns_ks_lengths.begin(), b_ns_ks_lengths.end()), std::vector<std::size_t>(b_gs_ns_ks_lengths.begin(), b_gs_ns_ks_lengths.end()),
std::vector<std::size_t>(b_ns_ks_strides.begin(), b_ns_ks_strides.end())); std::vector<std::size_t>(b_gs_ns_ks_strides.begin(), b_gs_ns_ks_strides.end()));
Tensor<DDataType> d_ms_ns( Tensor<DDataType> d_gs_ms_ns(
std::vector<std::size_t>(d_ms_ns_lengths.begin(), d_ms_ns_lengths.end()), std::vector<std::size_t>(d_gs_ms_ns_lengths.begin(), d_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(d_ms_ns_strides.begin(), d_ms_ns_strides.end())); std::vector<std::size_t>(d_gs_ms_ns_strides.begin(), d_gs_ms_ns_strides.end()));
Tensor<EDataType> e_ms_ns_host_result( Tensor<EDataType> e_gs_ms_ns_host_result(
std::vector<std::size_t>(e_ms_ns_lengths.begin(), e_ms_ns_lengths.end()), std::vector<std::size_t>(e_gs_ms_ns_lengths.begin(), e_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(e_ms_ns_strides.begin(), e_ms_ns_strides.end())); std::vector<std::size_t>(e_gs_ms_ns_strides.begin(), e_gs_ms_ns_strides.end()));
Tensor<EDataType> e_ms_ns_device_result( Tensor<EDataType> e_gs_ms_ns_device_result(
std::vector<std::size_t>(e_ms_ns_lengths.begin(), e_ms_ns_lengths.end()), std::vector<std::size_t>(e_gs_ms_ns_lengths.begin(), e_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(e_ms_ns_strides.begin(), e_ms_ns_strides.end())); std::vector<std::size_t>(e_gs_ms_ns_strides.begin(), e_gs_ms_ns_strides.end()));
std::cout << "a_ms_ks: " << a_ms_ks.mDesc << std::endl; std::cout << "a_gs_ms_ks: " << a_gs_ms_ks.mDesc << std::endl;
std::cout << "b_ns_ks: " << b_ns_ks.mDesc << std::endl; std::cout << "b_gs_ns_ks: " << b_gs_ns_ks.mDesc << std::endl;
std::cout << "d_ms_ns: " << d_ms_ns.mDesc << std::endl; std::cout << "d_gs_ms_ns: " << d_gs_ms_ns.mDesc << std::endl;
std::cout << "e_ms_ns: " << e_ms_ns_host_result.mDesc << std::endl; std::cout << "e_gs_ms_ns: " << e_gs_ms_ns_host_result.mDesc << std::endl;
switch(init_method) switch(init_method)
{ {
case 0: break; case 0: break;
case 1: case 1:
a_ms_ks.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5}); a_gs_ms_ks.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5});
b_ns_ks.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5}); b_gs_ns_ks.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
d_ms_ns.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5}); d_gs_ms_ns.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
break; break;
default: default:
a_ms_ks.GenerateTensorValue(GeneratorTensor_3<ADataType>{0.0, 1.0}); a_gs_ms_ks.GenerateTensorValue(GeneratorTensor_3<ADataType>{0.0, 1.0});
b_ns_ks.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5}); b_gs_ns_ks.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5});
d_ms_ns.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5}); d_gs_ms_ns.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5});
break; break;
} }
DeviceMem a_device_buf(sizeof(ADataType) * a_ms_ks.mDesc.GetElementSpaceSize()); DeviceMem a_device_buf(sizeof(ADataType) * a_gs_ms_ks.mDesc.GetElementSpaceSize());
DeviceMem b_device_buf(sizeof(BDataType) * b_ns_ks.mDesc.GetElementSpaceSize()); DeviceMem b_device_buf(sizeof(BDataType) * b_gs_ns_ks.mDesc.GetElementSpaceSize());
DeviceMem d_device_buf(sizeof(DDataType) * d_ms_ns.mDesc.GetElementSpaceSize()); DeviceMem d_device_buf(sizeof(DDataType) * d_gs_ms_ns.mDesc.GetElementSpaceSize());
DeviceMem e_device_buf(sizeof(EDataType) * e_ms_ns_device_result.mDesc.GetElementSpaceSize()); DeviceMem e_device_buf(sizeof(EDataType) *
e_gs_ms_ns_device_result.mDesc.GetElementSpaceSize());
a_device_buf.ToDevice(a_ms_ks.mData.data()); a_device_buf.ToDevice(a_gs_ms_ks.mData.data());
b_device_buf.ToDevice(b_ns_ks.mData.data()); b_device_buf.ToDevice(b_gs_ns_ks.mData.data());
d_device_buf.ToDevice(d_ms_ns.mData.data()); d_device_buf.ToDevice(d_gs_ms_ns.mData.data());
// set zero // set zero
e_device_buf.SetZero(); e_device_buf.SetZero();
...@@ -296,14 +306,14 @@ int main(int argc, char* argv[]) ...@@ -296,14 +306,14 @@ int main(int argc, char* argv[])
b_device_buf.GetDeviceBuffer(), b_device_buf.GetDeviceBuffer(),
std::array<const void*, 1>{d_device_buf.GetDeviceBuffer()}, std::array<const void*, 1>{d_device_buf.GetDeviceBuffer()},
e_device_buf.GetDeviceBuffer(), e_device_buf.GetDeviceBuffer(),
a_ms_ks_lengths, a_gs_ms_ks_lengths,
a_ms_ks_strides, a_gs_ms_ks_strides,
b_ns_ks_lengths, b_gs_ns_ks_lengths,
b_ns_ks_strides, b_gs_ns_ks_strides,
std::array<std::vector<ck::index_t>, 1>{d_ms_ns_lengths}, std::array<std::vector<ck::index_t>, 1>{d_gs_ms_ns_lengths},
std::array<std::vector<ck::index_t>, 1>{d_ms_ns_strides}, std::array<std::vector<ck::index_t>, 1>{d_gs_ms_ns_strides},
e_ms_ns_lengths, e_gs_ms_ns_lengths,
e_ms_ns_strides, e_gs_ms_ns_strides,
a_element_op, a_element_op,
b_element_op, b_element_op,
cde_element_op); cde_element_op);
...@@ -317,18 +327,18 @@ int main(int argc, char* argv[]) ...@@ -317,18 +327,18 @@ int main(int argc, char* argv[])
float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel}); float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});
ck::index_t M = std::accumulate(e_ms_ns_lengths.begin(), std::size_t M = std::accumulate(e_gs_ms_ns_lengths.begin() + NumDimG,
e_ms_ns_lengths.begin() + NumDimM, e_gs_ms_ns_lengths.begin() + NumDimG + NumDimM,
ck::index_t{1}, ck::index_t{1},
std::multiplies<ck::index_t>{}); std::multiplies<ck::index_t>{});
ck::index_t N = std::accumulate(e_ms_ns_lengths.begin() + NumDimM, std::size_t N = std::accumulate(e_gs_ms_ns_lengths.begin() + NumDimG + NumDimM,
e_ms_ns_lengths.begin() + NumDimM + NumDimN, e_gs_ms_ns_lengths.begin() + NumDimG + NumDimM + NumDimN,
ck::index_t{1}, ck::index_t{1},
std::multiplies<ck::index_t>{}); std::multiplies<ck::index_t>{});
ck::index_t K = std::accumulate(a_ms_ks_lengths.begin() + NumDimM, std::size_t K = std::accumulate(a_gs_ms_ks_lengths.begin() + NumDimG + NumDimM,
a_ms_ks_lengths.begin() + NumDimM + NumDimK, a_gs_ms_ks_lengths.begin() + NumDimG + NumDimM + NumDimK,
ck::index_t{1}, ck::index_t{1},
std::multiplies<ck::index_t>{}); std::multiplies<ck::index_t>{});
...@@ -343,53 +353,63 @@ int main(int argc, char* argv[]) ...@@ -343,53 +353,63 @@ int main(int argc, char* argv[])
std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s, " std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s, "
<< op.GetTypeString() << std::endl; << op.GetTypeString() << std::endl;
e_device_buf.FromDevice(e_ms_ns_device_result.mData.data()); e_device_buf.FromDevice(e_gs_ms_ns_device_result.mData.data());
if(do_verification) if(do_verification)
{ {
Tensor<CShuffleDataType> c_ms_ns_host_result( Tensor<CShuffleDataType> c_gs_ms_ns_host_result(
std::vector<std::size_t>(e_ms_ns_lengths.begin(), e_ms_ns_lengths.end()), std::vector<std::size_t>(e_gs_ms_ns_lengths.begin(), e_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(e_ms_ns_strides.begin(), e_ms_ns_strides.end())); std::vector<std::size_t>(e_gs_ms_ns_strides.begin(), e_gs_ms_ns_strides.end()));
using ReferenceOpInstance = ReferenceContraction_M2_N3_K1<NumDimM, using ReferenceOpInstance = ReferenceContraction_G1_M2_N3_K1<NumDimM,
NumDimN, NumDimN,
NumDimK, NumDimK,
ADataType, ADataType,
BDataType, BDataType,
CShuffleDataType, CShuffleDataType,
AccDataType, AccDataType,
AElementOp, AElementOp,
BElementOp, BElementOp,
PassThrough>; PassThrough>;
auto ref_gemm = ReferenceOpInstance{}; auto ref_gemm = ReferenceOpInstance{};
auto ref_invoker = ref_gemm.MakeInvoker(); auto ref_invoker = ref_gemm.MakeInvoker();
auto ref_argument = ref_gemm.MakeArgument( auto ref_argument = ref_gemm.MakeArgument(a_gs_ms_ks,
a_ms_ks, b_ns_ks, c_ms_ns_host_result, a_element_op, b_element_op, PassThrough{}); b_gs_ns_ks,
c_gs_ms_ns_host_result,
a_element_op,
b_element_op,
PassThrough{});
ref_invoker.Run(ref_argument); ref_invoker.Run(ref_argument);
for(size_t m0 = 0; m0 < e_ms_ns_host_result.mDesc.GetLengths()[0]; ++m0) for(size_t g0 = 0; g0 < e_gs_ms_ns_host_result.mDesc.GetLengths()[0]; ++g0)
{ {
for(size_t m1 = 0; m1 < e_ms_ns_host_result.mDesc.GetLengths()[1]; ++m1) for(size_t m0 = 0; m0 < e_gs_ms_ns_host_result.mDesc.GetLengths()[1]; ++m0)
{ {
for(size_t n0 = 0; n0 < e_ms_ns_host_result.mDesc.GetLengths()[2]; ++n0) for(size_t m1 = 0; m1 < e_gs_ms_ns_host_result.mDesc.GetLengths()[2]; ++m1)
{ {
for(size_t n1 = 0; n1 < e_ms_ns_host_result.mDesc.GetLengths()[3]; ++n1) for(size_t n0 = 0; n0 < e_gs_ms_ns_host_result.mDesc.GetLengths()[3]; ++n0)
{ {
for(size_t n2 = 0; n2 < e_ms_ns_host_result.mDesc.GetLengths()[4]; ++n2) for(size_t n1 = 0; n1 < e_gs_ms_ns_host_result.mDesc.GetLengths()[4]; ++n1)
{ {
cde_element_op(e_ms_ns_host_result(m0, m1, n0, n1, n2), for(size_t n2 = 0; n2 < e_gs_ms_ns_host_result.mDesc.GetLengths()[5];
c_ms_ns_host_result(m0, m1, n0, n1, n2), ++n2)
d_ms_ns(m0, m1, n0, n1, n2)); {
cde_element_op(e_gs_ms_ns_host_result(g0, m0, m1, n0, n1, n2),
c_gs_ms_ns_host_result(g0, m0, m1, n0, n1, n2),
d_gs_ms_ns(g0, m0, m1, n0, n1, n2));
}
} }
} }
} }
} }
} }
return ck::utils::check_err(e_ms_ns_device_result.mData, e_ms_ns_host_result.mData) ? 0 : 1; return ck::utils::check_err(e_gs_ms_ns_device_result.mData, e_gs_ms_ns_host_result.mData)
? 0
: 1;
} }
return 0; return 0;
......
example/25_gemm_bias_e_permute/gemm_bias_e_permute_m3n2_xdl_fp16.cpp → example/25_gemm_bias_e_permute/gemm_bias_e_permute_g1m3n2k1_xdl_fp16.cpp
View file @ 35e49f2d
...@@ -33,7 +33,7 @@ using DDataType = F16; ...@@ -33,7 +33,7 @@ using DDataType = F16;
using DsDataType = ck::Tuple<DDataType>; using DsDataType = ck::Tuple<DDataType>;
using EDataType = F16; using EDataType = F16;
static constexpr ck::index_t NumDimG = 0; static constexpr ck::index_t NumDimG = 1;
static constexpr ck::index_t NumDimM = 3; static constexpr ck::index_t NumDimM = 3;
static constexpr ck::index_t NumDimN = 2; static constexpr ck::index_t NumDimN = 2;
static constexpr ck::index_t NumDimK = 1; static constexpr ck::index_t NumDimK = 1;
...@@ -53,13 +53,13 @@ using DeviceOpInstanceKKNN = ck::tensor_operation::device:: ...@@ -53,13 +53,13 @@ using DeviceOpInstanceKKNN = ck::tensor_operation::device::
//############################################| | | | | Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Spacialization| Spacialization| Spacialization| Spacialization| Prefetch| Size| Block| Block| Block| | | XDL| XDL| Per| Per| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector| //############################################| | | | | Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Spacialization| Spacialization| Spacialization| Spacialization| Prefetch| Size| Block| Block| Block| | | XDL| XDL| Per| Per| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
//############################################| | | | | | | | | | | Operation| Operation| Operation| | | | | Stage| | | | | | | | | Wave| Wave| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl| //############################################| | | | | | | | | | | Operation| Operation| Operation| | | | | Stage| | | | | | | | | Wave| Wave| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
//############################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | //############################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
DeviceBatchedContractionMultipleD_Xdl_CShuffle< NumDimG, NumDimM, NumDimN, NumDimK, F16, F16, F32, F16, DsDataType, F16, AElementOp, BElementOp, CDEElementOp, GemmSpec, ABSpec, ABSpec, DESpec, 1, 256, 256, 128, 32, 8, 8, 32, 32, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 4>, 1>; DeviceBatchedContractionMultipleD_Xdl_CShuffle< NumDimG, NumDimM, NumDimN, NumDimK, F16, F16, F32, F16, DsDataType, F16, AElementOp, BElementOp, CDEElementOp, GemmSpec, ABSpec, ABSpec, DESpec, 1, 256, 256, 128, 32, 8, 8, 32, 32, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 4>, 8>;
// clang-format on // clang-format on
using DeviceOpInstance = DeviceOpInstanceKKNN; using DeviceOpInstance = DeviceOpInstanceKKNN;
// hardcoded for NumDimM == NumDimN == NumDimK == 2 template <ck::index_t NumDimG,
template <ck::index_t NumDimM, ck::index_t NumDimM,
ck::index_t NumDimN, ck::index_t NumDimN,
ck::index_t NumDimK, ck::index_t NumDimK,
typename ADataType, typename ADataType,
...@@ -69,30 +69,31 @@ template <ck::index_t NumDimM, ...@@ -69,30 +69,31 @@ template <ck::index_t NumDimM,
typename AElementwiseOperation, typename AElementwiseOperation,
typename BElementwiseOperation, typename BElementwiseOperation,
typename CDEElementwiseOperation, typename CDEElementwiseOperation,
ck::enable_if_t<NumDimM == 3 && NumDimN == 2 && NumDimK == 1, bool> = false> ck::enable_if_t<NumDimG == 1 && NumDimM == 3 && NumDimN == 2 && NumDimK == 1, bool> =
struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::BaseOperator false>
struct ReferenceContraction_G1_M3_N2_K1 : public ck::tensor_operation::device::BaseOperator
{ {
// Argument // Argument
struct Argument : public ck::tensor_operation::device::BaseArgument struct Argument : public ck::tensor_operation::device::BaseArgument
{ {
Argument(const Tensor<ADataType>& a_ms_ks, Argument(const Tensor<ADataType>& a_gs_ms_ks,
const Tensor<BDataType>& b_ns_ks, const Tensor<BDataType>& b_gs_ns_ks,
Tensor<EDataType>& e_ms_ns, Tensor<EDataType>& e_gs_ms_ns,
AElementwiseOperation a_element_op, AElementwiseOperation a_element_op,
BElementwiseOperation b_element_op, BElementwiseOperation b_element_op,
CDEElementwiseOperation cde_element_op) CDEElementwiseOperation cde_element_op)
: a_ms_ks_{a_ms_ks}, : a_gs_ms_ks_{a_gs_ms_ks},
b_ns_ks_{b_ns_ks}, b_gs_ns_ks_{b_gs_ns_ks},
e_ms_ns_{e_ms_ns}, e_gs_ms_ns_{e_gs_ms_ns},
a_element_op_{a_element_op}, a_element_op_{a_element_op},
b_element_op_{b_element_op}, b_element_op_{b_element_op},
cde_element_op_{cde_element_op} cde_element_op_{cde_element_op}
{ {
} }
const Tensor<ADataType>& a_ms_ks_; const Tensor<ADataType>& a_gs_ms_ks_;
const Tensor<BDataType>& b_ns_ks_; const Tensor<BDataType>& b_gs_ns_ks_;
Tensor<EDataType>& e_ms_ns_; Tensor<EDataType>& e_gs_ms_ns_;
AElementwiseOperation a_element_op_; AElementwiseOperation a_element_op_;
BElementwiseOperation b_element_op_; BElementwiseOperation b_element_op_;
...@@ -102,12 +103,12 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base ...@@ -102,12 +103,12 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base
// Invoker // Invoker
struct Invoker : public ck::tensor_operation::device::BaseInvoker struct Invoker : public ck::tensor_operation::device::BaseInvoker
{ {
using Argument = ReferenceContraction_M3_N2_K1::Argument; using Argument = ReferenceContraction_G1_M3_N2_K1::Argument;
float Run(const Argument& arg) float Run(const Argument& arg)
{ {
auto f_ms_ns = [&](auto m0, auto m1, auto m2, auto n0, auto n1) { auto f_gs_ms_ns = [&](auto g0, auto m0, auto m1, auto m2, auto n0, auto n1) {
const int K0 = arg.a_ms_ks_.mDesc.GetLengths()[3]; const int K0 = arg.a_gs_ms_ks_.mDesc.GetLengths()[4];
AccDataType v_acc = 0; AccDataType v_acc = 0;
...@@ -117,9 +118,10 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base ...@@ -117,9 +118,10 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base
AccDataType v_b; AccDataType v_b;
arg.a_element_op_( arg.a_element_op_(
v_a, ck::type_convert<const AccDataType>(arg.a_ms_ks_(m0, m1, m2, k0))); v_a,
ck::type_convert<const AccDataType>(arg.a_gs_ms_ks_(g0, m0, m1, m2, k0)));
arg.b_element_op_( arg.b_element_op_(
v_b, ck::type_convert<const AccDataType>(arg.b_ns_ks_(n0, n1, k0))); v_b, ck::type_convert<const AccDataType>(arg.b_gs_ns_ks_(g0, n0, n1, k0)));
v_acc += v_a * v_b; v_acc += v_a * v_b;
} }
...@@ -128,15 +130,16 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base ...@@ -128,15 +130,16 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base
arg.cde_element_op_(v_c, v_acc); arg.cde_element_op_(v_c, v_acc);
arg.e_ms_ns_(m0, m1, m2, n0, n1) = v_c; arg.e_gs_ms_ns_(g0, m0, m1, m2, n0, n1) = v_c;
}; };
make_ParallelTensorFunctor(f_ms_ns, make_ParallelTensorFunctor(f_gs_ms_ns,
arg.e_ms_ns_.mDesc.GetLengths()[0], arg.e_gs_ms_ns_.mDesc.GetLengths()[0],
arg.e_ms_ns_.mDesc.GetLengths()[1], arg.e_gs_ms_ns_.mDesc.GetLengths()[1],
arg.e_ms_ns_.mDesc.GetLengths()[2], arg.e_gs_ms_ns_.mDesc.GetLengths()[2],
arg.e_ms_ns_.mDesc.GetLengths()[3], arg.e_gs_ms_ns_.mDesc.GetLengths()[3],
arg.e_ms_ns_.mDesc.GetLengths()[4])( arg.e_gs_ms_ns_.mDesc.GetLengths()[4],
arg.e_gs_ms_ns_.mDesc.GetLengths()[5])(
std::thread::hardware_concurrency()); std::thread::hardware_concurrency());
return 0; return 0;
...@@ -160,14 +163,15 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base ...@@ -160,14 +163,15 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base
return true; return true;
} }
static auto MakeArgument(const Tensor<ADataType>& a_ms_ks, static auto MakeArgument(const Tensor<ADataType>& a_gs_ms_ks,
const Tensor<BDataType>& b_ns_ks, const Tensor<BDataType>& b_gs_ns_ks,
Tensor<EDataType>& e_ms_ns, Tensor<EDataType>& e_gs_ms_ns,
AElementwiseOperation a_element_op, AElementwiseOperation a_element_op,
BElementwiseOperation b_element_op, BElementwiseOperation b_element_op,
CDEElementwiseOperation cde_element_op) CDEElementwiseOperation cde_element_op)
{ {
return Argument{a_ms_ks, b_ns_ks, e_ms_ns, a_element_op, b_element_op, cde_element_op}; return Argument{
a_gs_ms_ks, b_gs_ns_ks, e_gs_ms_ns, a_element_op, b_element_op, cde_element_op};
} }
static auto MakeInvoker() { return Invoker{}; } static auto MakeInvoker() { return Invoker{}; }
...@@ -182,7 +186,7 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base ...@@ -182,7 +186,7 @@ struct ReferenceContraction_M3_N2_K1 : public ck::tensor_operation::device::Base
auto str = std::stringstream(); auto str = std::stringstream();
// clang-format off // clang-format off
str << "ReferenceContraction_M3_N2_K1" str << "ReferenceContraction_G1_M3_N2_K1"
<< std::endl; << std::endl;
// clang-format on // clang-format on
...@@ -196,36 +200,33 @@ int main(int argc, char* argv[]) ...@@ -196,36 +200,33 @@ int main(int argc, char* argv[])
int init_method = 1; int init_method = 1;
bool time_kernel = false; bool time_kernel = false;
ck::index_t G0 = 1;
ck::index_t M0 = 4; ck::index_t M0 = 4;
ck::index_t M1 = 32; ck::index_t M1 = 8;
ck::index_t M2 = 128; ck::index_t M2 = 256;
ck::index_t N0 = 16; ck::index_t N0 = 32;
ck::index_t N1 = 256; ck::index_t N1 = 128;
ck::index_t K0 = 256; ck::index_t K0 = 1024;
// A[M0, M1, M2, K0] // A[M0, M1, M2, K0]
std::vector<ck::index_t> a_ms_ks_lengths{M0, M1, M2, K0}; std::vector<ck::index_t> a_gs_ms_ks_lengths{G0, M0, M1, M2, K0};
std::vector<ck::index_t> a_ms_ks_strides{M1 * M2 * K0, M2 * K0, K0, 1}; std::vector<ck::index_t> a_gs_ms_ks_strides{M0 * M1 * M2 * K0, M1 * M2 * K0, M2 * K0, K0, 1};
// B[N0, N1, K0] // B[N0, N1, K0]
std::vector<ck::index_t> b_ns_ks_lengths{N0, N1, K0}; std::vector<ck::index_t> b_gs_ns_ks_lengths{G0, N0, N1, K0};
std::vector<ck::index_t> b_ns_ks_strides{N1 * K0, K0, 1}; std::vector<ck::index_t> b_gs_ns_ks_strides{N0 * N1 * K0, N1 * K0, K0, 1};
#if 1
// D[M0, N0, M1, N1, M2]
std::vector<ck::index_t> d_ms_ns_lengths{M0, M1, M2, N0, N1};
std::vector<ck::index_t> d_ms_ns_strides{0, 0, 0, N1, 1};
// E[M0, N0, M1, N1, M2]
std::vector<ck::index_t> e_ms_ns_lengths{M0, M1, M2, N0, N1};
std::vector<ck::index_t> e_ms_ns_strides{N0 * M1 * N1 * M2, N1 * M2, 1, M1 * N1 * M2, M2};
#else
// D[M0, N0, M1, N1, M2] // D[M0, N0, M1, N1, M2]
std::vector<ck::index_t> d_ms_ns_lengths{M0, M1, M2, N0, N1}; std::vector<ck::index_t> d_gs_ms_ns_lengths{G0, M0, M1, M2, N0, N1};
std::vector<ck::index_t> d_ms_ns_strides{0, 0, 0, N1, 1}; std::vector<ck::index_t> d_gs_ms_ns_strides{N0 * N1, 0, 0, 0, N1, 1};
// E[M0, N0, M1, N1, M2]
std::vector<ck::index_t> e_ms_ns_lengths{M0, M1, M2, N0, N1}; // E[M1, M0, N0, M1, N1]
std::vector<ck::index_t> e_ms_ns_strides{M1 * M2 * N0 * N1, M2 * N0 * N1, N0 * N1, N1, 1}; std::vector<ck::index_t> e_gs_ms_ns_lengths{G0, M0, M1, M2, N0, N1};
#endif std::vector<ck::index_t> e_gs_ms_ns_strides{
M0 * M1 * M2 * N1 * N0, N0 * M1 * N1, N1, M0 * N0 * M1 * N1, M1 * N1, 1};
if(argc == 1) if(argc == 1)
{ {
...@@ -245,50 +246,51 @@ int main(int argc, char* argv[]) ...@@ -245,50 +246,51 @@ int main(int argc, char* argv[])
exit(0); exit(0);
} }
Tensor<ADataType> a_ms_ks( Tensor<ADataType> a_gs_ms_ks(
std::vector<std::size_t>(a_ms_ks_lengths.begin(), a_ms_ks_lengths.end()), std::vector<std::size_t>(a_gs_ms_ks_lengths.begin(), a_gs_ms_ks_lengths.end()),
std::vector<std::size_t>(a_ms_ks_strides.begin(), a_ms_ks_strides.end())); std::vector<std::size_t>(a_gs_ms_ks_strides.begin(), a_gs_ms_ks_strides.end()));
Tensor<BDataType> b_ns_ks( Tensor<BDataType> b_gs_ns_ks(
std::vector<std::size_t>(b_ns_ks_lengths.begin(), b_ns_ks_lengths.end()), std::vector<std::size_t>(b_gs_ns_ks_lengths.begin(), b_gs_ns_ks_lengths.end()),
std::vector<std::size_t>(b_ns_ks_strides.begin(), b_ns_ks_strides.end())); std::vector<std::size_t>(b_gs_ns_ks_strides.begin(), b_gs_ns_ks_strides.end()));
Tensor<DDataType> d_ms_ns( Tensor<DDataType> d_gs_ms_ns(
std::vector<std::size_t>(d_ms_ns_lengths.begin(), d_ms_ns_lengths.end()), std::vector<std::size_t>(d_gs_ms_ns_lengths.begin(), d_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(d_ms_ns_strides.begin(), d_ms_ns_strides.end())); std::vector<std::size_t>(d_gs_ms_ns_strides.begin(), d_gs_ms_ns_strides.end()));
Tensor<EDataType> e_ms_ns_host_result( Tensor<EDataType> e_gs_ms_ns_host_result(
std::vector<std::size_t>(e_ms_ns_lengths.begin(), e_ms_ns_lengths.end()), std::vector<std::size_t>(e_gs_ms_ns_lengths.begin(), e_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(e_ms_ns_strides.begin(), e_ms_ns_strides.end())); std::vector<std::size_t>(e_gs_ms_ns_strides.begin(), e_gs_ms_ns_strides.end()));
Tensor<EDataType> e_ms_ns_device_result( Tensor<EDataType> e_gs_ms_ns_device_result(
std::vector<std::size_t>(e_ms_ns_lengths.begin(), e_ms_ns_lengths.end()), std::vector<std::size_t>(e_gs_ms_ns_lengths.begin(), e_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(e_ms_ns_strides.begin(), e_ms_ns_strides.end())); std::vector<std::size_t>(e_gs_ms_ns_strides.begin(), e_gs_ms_ns_strides.end()));
std::cout << "a_ms_ks: " << a_ms_ks.mDesc << std::endl; std::cout << "a_gs_ms_ks: " << a_gs_ms_ks.mDesc << std::endl;
std::cout << "b_ns_ks: " << b_ns_ks.mDesc << std::endl; std::cout << "b_gs_ns_ks: " << b_gs_ns_ks.mDesc << std::endl;
std::cout << "d_ms_ns: " << d_ms_ns.mDesc << std::endl; std::cout << "d_gs_ms_ns: " << d_gs_ms_ns.mDesc << std::endl;
std::cout << "e_ms_ns: " << e_ms_ns_host_result.mDesc << std::endl; std::cout << "e_gs_ms_ns: " << e_gs_ms_ns_host_result.mDesc << std::endl;
switch(init_method) switch(init_method)
{ {
case 0: break; case 0: break;
case 1: case 1:
a_ms_ks.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5}); a_gs_ms_ks.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5});
b_ns_ks.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5}); b_gs_ns_ks.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
d_ms_ns.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5}); d_gs_ms_ns.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
break; break;
default: default:
a_ms_ks.GenerateTensorValue(GeneratorTensor_3<ADataType>{0.0, 1.0}); a_gs_ms_ks.GenerateTensorValue(GeneratorTensor_3<ADataType>{0.0, 1.0});
b_ns_ks.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5}); b_gs_ns_ks.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5});
d_ms_ns.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5}); d_gs_ms_ns.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5});
break; break;
} }
DeviceMem a_device_buf(sizeof(ADataType) * a_ms_ks.mDesc.GetElementSpaceSize()); DeviceMem a_device_buf(sizeof(ADataType) * a_gs_ms_ks.mDesc.GetElementSpaceSize());
DeviceMem b_device_buf(sizeof(BDataType) * b_ns_ks.mDesc.GetElementSpaceSize()); DeviceMem b_device_buf(sizeof(BDataType) * b_gs_ns_ks.mDesc.GetElementSpaceSize());
DeviceMem d_device_buf(sizeof(DDataType) * d_ms_ns.mDesc.GetElementSpaceSize()); DeviceMem d_device_buf(sizeof(DDataType) * d_gs_ms_ns.mDesc.GetElementSpaceSize());
DeviceMem e_device_buf(sizeof(EDataType) * e_ms_ns_device_result.mDesc.GetElementSpaceSize()); DeviceMem e_device_buf(sizeof(EDataType) *
e_gs_ms_ns_device_result.mDesc.GetElementSpaceSize());
a_device_buf.ToDevice(a_ms_ks.mData.data()); a_device_buf.ToDevice(a_gs_ms_ks.mData.data());
b_device_buf.ToDevice(b_ns_ks.mData.data()); b_device_buf.ToDevice(b_gs_ns_ks.mData.data());
d_device_buf.ToDevice(d_ms_ns.mData.data()); d_device_buf.ToDevice(d_gs_ms_ns.mData.data());
// set zero // set zero
e_device_buf.SetZero(); e_device_buf.SetZero();
...@@ -304,14 +306,14 @@ int main(int argc, char* argv[]) ...@@ -304,14 +306,14 @@ int main(int argc, char* argv[])
b_device_buf.GetDeviceBuffer(), b_device_buf.GetDeviceBuffer(),
std::array<const void*, 1>{d_device_buf.GetDeviceBuffer()}, std::array<const void*, 1>{d_device_buf.GetDeviceBuffer()},
e_device_buf.GetDeviceBuffer(), e_device_buf.GetDeviceBuffer(),
a_ms_ks_lengths, a_gs_ms_ks_lengths,
a_ms_ks_strides, a_gs_ms_ks_strides,
b_ns_ks_lengths, b_gs_ns_ks_lengths,
b_ns_ks_strides, b_gs_ns_ks_strides,
std::array<std::vector<ck::index_t>, 1>{d_ms_ns_lengths}, std::array<std::vector<ck::index_t>, 1>{d_gs_ms_ns_lengths},
std::array<std::vector<ck::index_t>, 1>{d_ms_ns_strides}, std::array<std::vector<ck::index_t>, 1>{d_gs_ms_ns_strides},
e_ms_ns_lengths, e_gs_ms_ns_lengths,
e_ms_ns_strides, e_gs_ms_ns_strides,
a_element_op, a_element_op,
b_element_op, b_element_op,
cde_element_op); cde_element_op);
...@@ -325,18 +327,18 @@ int main(int argc, char* argv[]) ...@@ -325,18 +327,18 @@ int main(int argc, char* argv[])
float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel}); float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});
ck::index_t M = std::accumulate(e_ms_ns_lengths.begin(), ck::index_t M = std::accumulate(e_gs_ms_ns_lengths.begin(),
e_ms_ns_lengths.begin() + NumDimM, e_gs_ms_ns_lengths.begin() + NumDimM,
ck::index_t{1}, ck::index_t{1},
std::multiplies<ck::index_t>{}); std::multiplies<ck::index_t>{});
ck::index_t N = std::accumulate(e_ms_ns_lengths.begin() + NumDimM, ck::index_t N = std::accumulate(e_gs_ms_ns_lengths.begin() + NumDimM,
e_ms_ns_lengths.begin() + NumDimM + NumDimN, e_gs_ms_ns_lengths.begin() + NumDimM + NumDimN,
ck::index_t{1}, ck::index_t{1},
std::multiplies<ck::index_t>{}); std::multiplies<ck::index_t>{});
ck::index_t K = std::accumulate(a_ms_ks_lengths.begin() + NumDimM, ck::index_t K = std::accumulate(a_gs_ms_ks_lengths.begin() + NumDimM,
a_ms_ks_lengths.begin() + NumDimM + NumDimK, a_gs_ms_ks_lengths.begin() + NumDimM + NumDimK,
ck::index_t{1}, ck::index_t{1},
std::multiplies<ck::index_t>{}); std::multiplies<ck::index_t>{});
...@@ -351,53 +353,64 @@ int main(int argc, char* argv[]) ...@@ -351,53 +353,64 @@ int main(int argc, char* argv[])
std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s, " std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s, "
<< op.GetTypeString() << std::endl; << op.GetTypeString() << std::endl;
e_device_buf.FromDevice(e_ms_ns_device_result.mData.data()); e_device_buf.FromDevice(e_gs_ms_ns_device_result.mData.data());
if(do_verification) if(do_verification)
{ {
Tensor<CShuffleDataType> c_ms_ns_host_result( Tensor<CShuffleDataType> c_gs_ms_ns_host_result(
std::vector<std::size_t>(e_ms_ns_lengths.begin(), e_ms_ns_lengths.end()), std::vector<std::size_t>(e_gs_ms_ns_lengths.begin(), e_gs_ms_ns_lengths.end()),
std::vector<std::size_t>(e_ms_ns_strides.begin(), e_ms_ns_strides.end())); std::vector<std::size_t>(e_gs_ms_ns_strides.begin(), e_gs_ms_ns_strides.end()));
using ReferenceOpInstance = ReferenceContraction_M3_N2_K1<NumDimM, using ReferenceOpInstance = ReferenceContraction_G1_M3_N2_K1<NumDimG,
NumDimN, NumDimM,
NumDimK, NumDimN,
ADataType, NumDimK,
BDataType, ADataType,
CShuffleDataType, BDataType,
AccDataType, CShuffleDataType,
AElementOp, AccDataType,
BElementOp, AElementOp,
PassThrough>; BElementOp,
PassThrough>;
auto ref_gemm = ReferenceOpInstance{}; auto ref_gemm = ReferenceOpInstance{};
auto ref_invoker = ref_gemm.MakeInvoker(); auto ref_invoker = ref_gemm.MakeInvoker();
auto ref_argument = ref_gemm.MakeArgument( auto ref_argument = ref_gemm.MakeArgument(a_gs_ms_ks,
a_ms_ks, b_ns_ks, c_ms_ns_host_result, a_element_op, b_element_op, PassThrough{}); b_gs_ns_ks,
c_gs_ms_ns_host_result,
a_element_op,
b_element_op,
PassThrough{});
ref_invoker.Run(ref_argument); ref_invoker.Run(ref_argument);
for(size_t m0 = 0; m0 < e_ms_ns_host_result.mDesc.GetLengths()[0]; ++m0) for(size_t g0 = 0; g0 < e_gs_ms_ns_host_result.mDesc.GetLengths()[0]; ++g0)
{ {
for(size_t m1 = 0; m1 < e_ms_ns_host_result.mDesc.GetLengths()[1]; ++m1) for(size_t m0 = 0; m0 < e_gs_ms_ns_host_result.mDesc.GetLengths()[1]; ++m0)
{ {
for(size_t m2 = 0; m2 < e_ms_ns_host_result.mDesc.GetLengths()[2]; ++m2) for(size_t m1 = 0; m1 < e_gs_ms_ns_host_result.mDesc.GetLengths()[2]; ++m1)
{ {
for(size_t n0 = 0; n0 < e_ms_ns_host_result.mDesc.GetLengths()[3]; ++n0) for(size_t m2 = 0; m2 < e_gs_ms_ns_host_result.mDesc.GetLengths()[3]; ++m2)
{ {
for(size_t n1 = 0; n1 < e_ms_ns_host_result.mDesc.GetLengths()[4]; ++n1) for(size_t n0 = 0; n0 < e_gs_ms_ns_host_result.mDesc.GetLengths()[4]; ++n0)
{ {
cde_element_op(e_ms_ns_host_result(m0, m1, m2, n0, n1), for(size_t n1 = 0; n1 < e_gs_ms_ns_host_result.mDesc.GetLengths()[5];
c_ms_ns_host_result(m0, m1, m2, n0, n1), ++n1)
d_ms_ns(m0, m1, m2, n0, n1)); {
cde_element_op(e_gs_ms_ns_host_result(g0, m0, m1, m2, n0, n1),
c_gs_ms_ns_host_result(g0, m0, m1, m2, n0, n1),
d_gs_ms_ns(g0, m0, m1, m2, n0, n1));
}
} }
} }
} }
} }
} }
return ck::utils::check_err(e_ms_ns_device_result.mData, e_ms_ns_host_result.mData) ? 0 : 1; return ck::utils::check_err(e_gs_ms_ns_device_result.mData, e_gs_ms_ns_host_result.mData)
? 0
: 1;
} }
return 0; return 0;
......
include/ck/tensor_operation/gpu/device/device_batched_contraction_multiple_d_xdl_cshuffle.hpp
View file @ 35e49f2d
...@@ -500,11 +500,8 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle ...@@ -500,11 +500,8 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
std::array<long_index_t, NumDTensor> ds_offset; std::array<long_index_t, NumDTensor> ds_offset;
static_for<0, NumDTensor, 1>{}([&](auto i) { static_for<0, NumDTensor, 1>{}([&](auto i) {
if constexpr(NumDimG > 0) ds_offset[i] =
ds_offset[i] = ds_grid_desc_g_m_n_[i].CalculateOffset(make_multi_index(g_idx, 0, 0));
ds_grid_desc_g_m_n_[i].CalculateOffset(make_multi_index(g_idx, 0, 0));
else
ds_offset[i] = 0;
}); });
return ds_offset; return ds_offset;
...@@ -512,10 +509,7 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle ...@@ -512,10 +509,7 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
__host__ __device__ constexpr long_index_t GetEPtrOffset(index_t g_idx) const __host__ __device__ constexpr long_index_t GetEPtrOffset(index_t g_idx) const
{ {
if constexpr(NumDimG > 0) return e_grid_desc_g_m_n_.CalculateOffset(make_multi_index(g_idx, 0, 0));
return e_grid_desc_g_m_n_.CalculateOffset(make_multi_index(g_idx, 0, 0));
else
return 0;
} }
private: private:
...@@ -634,6 +628,8 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle ...@@ -634,6 +628,8 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
compute_ptr_offset_of_batch_{ compute_ptr_offset_of_batch_{
a_batch_stride_, b_batch_stride_, ds_grid_desc_g_m_n_, e_grid_desc_g_m_n_} a_batch_stride_, b_batch_stride_, ds_grid_desc_g_m_n_, e_grid_desc_g_m_n_}
{ {
static_assert(NumDimG > 0 && NumDimM > 0 && NumDimN > 0 && NumDimK > 0, "");
// populate pointer, batch stride, desc for Ds // populate pointer, batch stride, desc for Ds
static_for<0, NumDTensor, 1>{}([&](auto i) { static_for<0, NumDTensor, 1>{}([&](auto i) {
using DDataType = remove_cvref_t<tuple_element_t<i.value, DsDataType>>; using DDataType = remove_cvref_t<tuple_element_t<i.value, DsDataType>>;
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment