TensorOps kernels refactoring

src/CMakeLists.txt

@@ -337,11 +337,13 @@ set( MIOpen_Source
     solver/softmarginloss/forward_softmarginloss.cpp
     solver/softmax/attn_softmax.cpp
     solver/softmax/softmax.cpp
+    solver/tensor/Op1dTensorGeneric.cpp
     subbuffers.cpp
     t5layernorm_api.cpp
     target_properties.cpp
     temp_file.cpp
     tensor.cpp
+    tensor/problem_description.cpp
     tensor_api.cpp
     transformers_adam_w_api.cpp
     seq_tensor.cpp

src/include/miopen/tensor/invoke_params.hpp

@@ -0,0 +1,97 @@
+/*******************************************************************************
+ *
+ * MIT License
+ *
+ * Copyright (c) 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ *******************************************************************************/
+
+#pragma once
+
+#include <miopen/invoke_params.hpp>
+#include <miopen/tensor.hpp>
+
+namespace miopen {
+
+namespace tensor {
+
+struct InvokeParams : public miopen::InvokeParams
+{
+    InvokeParams(miopenTensorOp_t tensorOp_,
+                 const void* alpha0_,
+                 const TensorDescriptor& aTensorDesc_,
+                 ConstData_t ATensor_,
+                 const void* alpha1_,
+                 const TensorDescriptor& bTensorDesc_,
+                 ConstData_t BTensor_,
+                 const void* beta_,
+                 const TensorDescriptor& cTensorDesc_,
+                 Data_t CTensor_,
+                 const size_t Aoffset_,
+                 const size_t Boffset_,
+                 const size_t Coffset_,
+                 const bool nonStandardSquash_)
+        : alpha0(alpha0_),
+          alpha1(alpha1_),
+          beta(beta_),
+          tensorOp(tensorOp_),
+          aTensorDesc(aTensorDesc_),
+          ATensor(ATensor_),
+          bTensorDesc(bTensorDesc_),
+          BTensor(BTensor_),
+          cTensorDesc(cTensorDesc_),
+          CTensor(CTensor_),
+          Aoffset(Aoffset_),
+          Boffset(Boffset_),
+          Coffset(Coffset_),
+          nonStandardSquash(nonStandardSquash_)
+    {
+    }
+
+    size_t GetWorkspaceSize() const { return 0; }
+    Data_t GetWorkspace() const { return nullptr; }
+
+public:
+    const void* alpha0;
+    const void* alpha1;
+    const void* beta;
+
+    miopenTensorOp_t tensorOp;
+
+    TensorDescriptor aTensorDesc;
+    ConstData_t ATensor;
+
+    TensorDescriptor bTensorDesc;
+    ConstData_t BTensor;
+
+    TensorDescriptor cTensorDesc;
+    Data_t CTensor;
+
+    size_t Aoffset;
+    size_t Boffset;
+    size_t Coffset;
+
+    bool nonStandardSquash;
+};
+
+} // namespace tensor
+
+} // namespace miopen

src/include/miopen/tensor/problem_description.hpp

@@ -0,0 +1,152 @@
+/*******************************************************************************
+ *
+ * MIT License
+ *
+ * Copyright (c) 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ *******************************************************************************/
+
+#pragma once
+
+#include <miopen/problem_description_base.hpp>
+#include <miopen/tensor.hpp>
+
+namespace miopen {
+
+struct NetworkConfig;
+
+namespace tensor {
+
+struct ProblemDescription : ProblemDescriptionBase
+{
+    ProblemDescription(const miopenTensorOp_t tensorOp_,
+                       const void* alpha0_,
+                       const void* alpha1_,
+                       const void* beta_,
+                       const TensorDescriptor& aTensorDesc_,
+                       const TensorDescriptor& bTensorDesc_,
+                       const TensorDescriptor& cTensorDesc_,
+                       const bool nonStandardSquash_)
+        : tensorOp(tensorOp_),
+          aTensorDesc(aTensorDesc_),
+          bTensorDesc(bTensorDesc_),
+          cTensorDesc(cTensorDesc_),
+          nonStandardSquash(nonStandardSquash_)
+    {
+        CheckAndAssignAlphaBeta(alpha0_, alpha1_, beta_);
+
+        if(aTensorDesc.GetElementSize() != cTensorDesc.GetElementSize())
+        {
+            MIOPEN_THROW("A and C Tensors do not match");
+        }
+
+        if(bTensorDesc.GetType() != cTensorDesc.GetType())
+        {
+            MIOPEN_THROW("Datatypes for B and C tensors do not match !");
+        }
+
+        auto blens = bTensorDesc.GetLengths();
+        auto clens = cTensorDesc.GetLengths();
+        if(clens.size() > 5)
+        {
+            MIOPEN_THROW("Tensor dimension larger than 5: " + std::to_string(clens.size()));
+        }
+
+        if(blens.size() != clens.size())
+        {
+            MIOPEN_THROW("Number of dims in B and C Tensors do not match: " +
+                         std::to_string(blens.size()) + ", " + std::to_string(clens.size()));
+        }
+
+        if(!nonStandardSquash)
+        {
+            for(std::size_t i = 0; i < clens.size(); i++)
+            {
+                if(blens[i] != 1 && blens[i] != clens[i])
+                {
+                    MIOPEN_THROW("BTensor dim != 1 && BTensor dim != CTensor dim: " +
+                                 std::to_string(i));
+                }
+            }
+        }
+        else
+        {
+            // non standard behavior because blens[1] can be not equalt to clens[1]
+            if(!(clens.size() == 3 && blens[0] == 1 && clens[0] == 1 && blens[2] == clens[2]))
+            {
+                MIOPEN_THROW(
+                    "Non standard squashed operation supported only for 3d tensors and for "
+                    "the specific configuration");
+            }
+        }
+    }
+
+    const miopenTensorOp_t GetTensorOp() const { return tensorOp; }
+
+    const void* GetAlpha0() const { return alpha0; }
+    const void* GetAlpha1() const { return alpha1; }
+    const void* GetBeta() const { return beta; }
+
+    const TensorDescriptor& GetATensorDesc() const { return aTensorDesc; }
+    const TensorDescriptor& GetBTensorDesc() const { return bTensorDesc; }
+    const TensorDescriptor& GetCTensorDesc() const { return cTensorDesc; }
+
+    const bool GetNonStandardSquash() const { return nonStandardSquash; }
+
+    NetworkConfig MakeNetworkConfig() const override;
+
+private:
+    void CheckAndAssignAlphaBeta(const void* alpha0_, const void* alpha1_, const void* beta_)
+    {
+        if(alpha0_ == nullptr)
+        {
+            MIOPEN_THROW(miopenStatusBadParm, "Alpha0 value is nullptr");
+        }
+        if(alpha1_ == nullptr)
+        {
+            MIOPEN_THROW(miopenStatusBadParm, "Alpha1 value is nullptr");
+        }
+        if(beta_ == nullptr)
+        {
+            MIOPEN_THROW(miopenStatusBadParm, "Beta value is nullptr");
+        }
+
+        alpha0 = alpha0_;
+        alpha1 = alpha1_;
+        beta   = beta_;
+    }
+
+    const miopenTensorOp_t tensorOp;
+
+    const void* alpha0;
+    const void* alpha1;
+    const void* beta;
+
+    const TensorDescriptor& aTensorDesc;
+    const TensorDescriptor& bTensorDesc;
+    const TensorDescriptor& cTensorDesc;
+
+    const bool nonStandardSquash;
+};
+
+} // namespace tensor
+
+} // namespace miopen

src/include/miopen/tensor/solvers.hpp

@@ -0,0 +1,62 @@
+/*******************************************************************************
+ *
+ * MIT License
+ *
+ * Copyright (c) 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ *******************************************************************************/
+
+#pragma once
+
+#include <miopen/solver.hpp>
+#include <miopen/tensor/problem_description.hpp>
+
+#include <utility>
+
+namespace miopen {
+
+namespace solver {
+
+namespace tensor {
+
+using TensorOpSolver = NonTunableSolverBase<ExecutionContext, miopen::tensor::ProblemDescription>;
+
+struct Op1dTensorGeneric final : TensorOpSolver
+{
+    const std::string& SolverDbId() const override { return GetSolverDbId<Op1dTensorGeneric>(); }
+
+    bool IsApplicable(const ExecutionContext& context,
+                      const miopen::tensor::ProblemDescription& problem) const override;
+
+    ConvSolution GetSolution(const ExecutionContext& context,
+                             const miopen::tensor::ProblemDescription& problem) const override;
+
+    std::size_t GetWorkspaceSize(const ExecutionContext& context,
+                                 const miopen::tensor::ProblemDescription& problem) const override;
+
+    bool MayNeedWorkspace() const override { return false; }
+};
+
+} // namespace tensor
+
+} // namespace solver
+
+} // namespace miopen

src/include/miopen/tensor_ops.hpp

@@ -189,6 +189,22 @@ MIOPEN_INTERNALS_EXPORT void OpTensor(const Handle& handle,
                                       size_t Coffset         = 0,
                                       bool nonStandardSquash = false);
 
+MIOPEN_INTERNALS_EXPORT void OpTensorNew(Handle& handle,
+                                         miopenTensorOp_t tensorOp,
+                                         const void* alpha0,
+                                         const TensorDescriptor& aTensorDesc,
+                                         ConstData_t ATensor,
+                                         const void* alpha1,
+                                         const TensorDescriptor& bTensorDesc,
+                                         ConstData_t BTensor,
+                                         const void* beta,
+                                         const TensorDescriptor& cTensorDesc,
+                                         Data_t CTensor,
+                                         size_t Aoffset         = 0,
+                                         size_t Boffset         = 0,
+                                         size_t Coffset         = 0,
+                                         bool nonStandardSquash = false);
+
 MIOPEN_INTERNALS_EXPORT void CopyTensor(const Handle& handle,
                                         const TensorDescriptor& srcDesc,
                                         ConstData_t src,