compute_primitives.h

// Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

#pragma once

#ifdef PADDLE_WITH_CUDA
#include <cuda_fp16.h>
#endif
#ifdef PADDLE_WITH_HIP
#include <hip/hip_fp16.h>
#endif

#include "paddle/fluid/platform/cuda_device_function.h"
#include "paddle/fluid/platform/float16.h"

namespace paddle {
namespace operators {
namespace kernel_primitives {
namespace details {

#ifdef __HIPCC__
constexpr int kReduceMaxThread = 256;
constexpr int kWarpSize = 64;
#else
constexpr int kReduceMaxThread = 128;
constexpr int kWarpSize = 32;
#endif

// kGlobalMode: block reduce, each block gets an output;
// kLocalMode: thread reduce, each thread gets an output;
enum ReduceMode { kGlobalMode, kLocalMode };

template <typename T>
class MPTypeTrait {
 public:
  using Type = T;
};

template <>
class MPTypeTrait<platform::float16> {
 public:
  using Type = float;
};

/**
 * @brief Will be used in BlockYReduce, get the index of reduce_num in shared
 * memory.
 */
__device__ __forceinline__ int SharedMemoryIndex(int index) {
  return (threadIdx.y + index) * blockDim.x + threadIdx.x;
}

template <typename T, typename ReduceOp>
__device__ __forceinline__ T WarpReduce(T val, ReduceOp reducer) {
  unsigned mask = 0u;
  CREATE_SHFL_MASK(mask, true);
  for (int stride = details::kWarpSize / 2; stride > 0; stride >>= 1) {
    T temp = paddle::platform::CudaShuffleDownSync(mask, val, stride);
    val = reducer(val, temp);
  }
  return val;
}

/* e.g.
 * |---------block---------|
 * |warp0|warp1|warp2|warp3|
 * |0~31|32~63|64~95|96~127|  ---->blockDim.x = 128
 *  \|/  \|/   \|/    \|/     ---->1. First WarpReduce in each warp
 * res0  res1  res2  res3     ---->2. Store result of each warp to shared memory
 *   \    \    /     /        ---->3. Load the result above from shared memory
 *        res                         to warp0 and process the second WarpReduce
 */

/**
 * @brief BlockXReduce reduce along blockDim.x.
 */
template <typename T, typename ReduceOp>
__device__ __forceinline__ T BlockXReduce(T val, ReduceOp reducer) {
  __syncthreads();
  using details::kWarpSize;
  __shared__ T shared[2 * kWarpSize];
  int block_dim_x = blockDim.x;
  if (blockDim.x > kWarpSize) {
    block_dim_x = blockDim.x / kWarpSize;
    int lane = threadIdx.x % kWarpSize;
    int tid = threadIdx.y * blockDim.x + threadIdx.x;
    int wid = tid / kWarpSize;
    int bid = threadIdx.y;
    val = WarpReduce(val, reducer);
    if (lane == 0) {
      shared[wid] = val;
    }
    __syncthreads();
    val = shared[bid * block_dim_x + lane];
  }

  unsigned mask = 0u;
  CREATE_SHFL_MASK(mask, true);
  for (int stride = 1; stride < block_dim_x; stride <<= 1) {
    T temp = paddle::platform::CudaShuffleDownSync(mask, val, stride);
    val = reducer(val, temp);
  }
  return val;
}

/**
 * @brief BlockYReduce reduce along blockDim.y.
 */
template <typename T, typename ReduceOp>
__device__ __forceinline__ T BlockYReduce(T val, ReduceOp reducer) {
  __shared__ T shared_memory[details::kReduceMaxThread];
  shared_memory[SharedMemoryIndex(0)] = val;
  for (int stride = blockDim.y / 2; stride > 0; stride >>= 1) {
    __syncthreads();
    if (threadIdx.y < stride && threadIdx.y + stride < blockDim.y) {
      T temp = shared_memory[SharedMemoryIndex(stride)];
      val = reducer(val, temp);
    }
    shared_memory[SharedMemoryIndex(0)] = val;
  }
  return val;
}

}  // namespace details

/**
 * @brief Perform unary calculation according to OpFunc. Size of input and
 * output are the same.
 *
 * @template paraments
 * InT: Data type of in.
 * OutT: Data type of out.
 * NX: The number of data columns loaded by each thread.
 * NY: The number of data rows loaded by each thread.
 * BlockSize: Identifies the current device thread index method. For GPU,
 * threadIdx.x is used as the thread index, and for xpu, core_id() is used as
 * the index. Currently only GPU was supported.
 * OpFunc: Compute functor which has an operator() as following:
 *     template <typename InT, typename OutT>
 *     struct XxxFunctor {
 *       HOSTDEVICE OutT operator()(const InT& a) const {
 *         return ...;
 *       }
 *     };
 *
 * @param：
 * out: The register pointer of out, the size is NX * NY.
 * in: The register pointer of in, the size is NX * NY.
 * compute: Compute function which was declared like OpFunc<InT, OutT>().
 */
template <typename InT, typename OutT, int NX, int NY, int BlockSize,
          class OpFunc>
__device__ __forceinline__ void ElementwiseUnary(OutT* out, const InT* in,
                                                 OpFunc compute) {
#pragma unroll
  for (int idx = 0; idx < NX * NY; idx++) {
    out[idx] = static_cast<OutT>(compute(in[idx]));
  }
}

/**
 * @brief Binary calculation according to OpFunc. Size of The input and output
 * are the same.
 *
 * @template paraments
 * InT: Data type of in1 and in2.
 * OutT: Data type of out.
 * NX: The number of data columns loaded by each thread.
 * NY: The number of data rows loaded by each thread.
 * BlockSize: Identifies the current device thread index method. For GPU,
 * threadIdx.x is used as the thread index, and for xpu, core_id() is used as
 * the index. Currently only GPU was supported.
 * OpFunc: Compute functor which has an operator() as following:
 *     template <typename InT, typename OutT>
 *     struct XxxFunctor {
 *       HOSTDEVICE OutT operator()(const InT& a, const InT& b) const {
 *         return ...;
 *       }
 *     };
 *
 * @param：
 * out: The register pointer of out, the size is NX * NY.
 * in1: The register pointer of fist input, size is NX * NY.
 * in2: The register pointer of second input, size is NX * NY.
 * compute: Compute function which was declared like OpFunc<InT, OutT>().
 */
template <typename InT, typename OutT, int NX, int NY, int BlockSize,
          class OpFunc>
__device__ __forceinline__ void ElementwiseBinary(OutT* out, const InT* in1,
                                                  const InT* in2,
                                                  OpFunc compute) {
#pragma unroll
  for (int idx = 0; idx < NX * NY; ++idx) {
    out[idx] = static_cast<OutT>(compute(in1[idx], in2[idx]));
  }
}

/**
 * @brief Ternary calculation according to OpFunc. Size of input and output
 * are the same.
 *
 * @template paraments
 * InT: Data type of in1 and in2.
 * OutT: Data type of out.
 * NX: The number of data columns loaded by each thread.
 * NY: The number of data rows loaded by each thread.
 * BlockSize: Identifies the current device thread index method. For GPU,
 * threadIdx.x is used as the thread index, and for xpu, core_id() is used as
 * the index. Currently only GPU was supported.
 * OpFunc: Compute functor which has an operator() as following
 *     template <typename InT, typename OutT>
 *     struct XxxFunctor {
 *       HOSTDEVICE OutT operator()(const InT& a, const InT& b, const InT& c)
 * const {
 *         return ...;
 *       }
 *     };
 *
 * @param
 * out: The register pointer of out, the size is NX * NY.
 * in1: The register pointer of fist input, size is NX * NY.
 * in2: The register pointer of second input, size is NX * NY.
 * in3: The register pointer of third input, size is NX * NY.
 * compute: Compute function which was declared like OpFunc<InT, OutT>().
 */
template <typename InT, typename OutT, int NX, int NY, int BlockSize,
          class OpFunc>
__device__ __forceinline__ void ElementwiseTernary(OutT* out, const InT* in1,
                                                   const InT* in2,
                                                   const InT* in3,
                                                   OpFunc compute) {
#pragma unroll
  for (int idx = 0; idx < NX * NY; ++idx) {
    out[idx] = static_cast<OutT>(compute(in1[idx], in2[idx], in3[idx]));
  }
}

/**
 * @brief Multivariate calculation according to OpFunc. Size of input and output
 * are the same.
 *
 * @template paraments
 * InT: Data type of in1 and in2.
 * OutT: Data type of out.
 * NX: The number of data columns loaded by each thread.
 * NY: The number of data rows loaded by each thread.
 * BlockSize: Identifies the current device thread index method. For GPU,
 * threadIdx.x is used as the thread index, and for xpu, core_id() is used as
 * the index. Currently only GPU was supported.
 * Arity: The size of ins
 * OpFunc: Compute functor which has an operator() as following:
 *     template <typename InT, typename OutT>
 *     struct XxxFunctor {
 *       HOSTDEVICE OutT operator()(const InT* args) const {
 *         return ...;
 *       }
 *     };
 *
 * @param
 * out: The register pointer of out, the size is NX * NY.
 * ins: An array of pointers consisting of multiple inputs.
 * compute: Compute function which was declared like OpFunc<InT, OutT>().
 */
template <typename InT, typename OutT, int NX, int NY, int BlockSize, int Arity,
          class OpFunc>
__device__ __forceinline__ void ElementwiseAny(OutT* out, InT (*ins)[NX * NY],
                                               OpFunc compute) {
  InT args[Arity];
#pragma unroll
  for (int idx = 0; idx < NX * NY; ++idx) {
#pragma unroll
    for (int j = 0; j < Arity; ++j) {
      args[j] = ins[j][idx];
    }
    out[idx] = static_cast<OutT>(compute(args));
  }
}

/**
 * @brief Binary calculation according to OpFunc. Shape of in1 and in2 are the
 * different. Shape of in1 is [1, NX], but in2's shape is [NY, NX], the output
 * shape is [NY, NX].
 *
 * @template paraments
 * InT: Data type of in1 and in2.
 * OutT: Data type of out.
 * NX: The number of data columns loaded by each thread.
 * NY: The number of data rows loaded by each thread.
 * BlockSize: Identifies the current device thread index method. For GPU,
 * threadIdx.x is used as the thread index, and for xpu, core_id() is used as
 * the index. Currently only GPU was supported.
 * OpFunc: Compute functor which has an operator() as following
 *     template <typename InT, typename OutT>
 *     struct XxxFunctor {
 *       HOSTDEVICE OutT operator()(const InT& a, const InT& b) const {
 *         return ...;
 *       }
 *     };
 *
 * @param
 * out: The register pointer of out, the size is NX * NY.
 * in1: The register pointer of fist input, size is NX * 1.
 * in2: The register pointer of second input, size is NX * NY.
 * compute: Compute function which was declared like OpFunc<InT, OutT>().
 */
template <typename T, typename OutT, int NX, int NY, int BlockSize,
          class OpFunc>
__device__ __forceinline__ void CycleBinary(OutT* out, const T* in1,
                                            const T* in2, OpFunc compute) {
#pragma unroll
  for (int idx = 0; idx < NX; idx++) {
#pragma unroll
    for (int idy = 0; idy < NY; idy++) {
      out[idx + idy * NX] =
          static_cast<OutT>(compute(in1[idx], in2[idx + idy * NX]));
    }
  }
}

/**
 * @brief The Reduce provides collective methods for computing a parallel
 * reduction of items partitioned across a CUDA block and intra thread. When
 * ReduceMode == kLocalMode, thread reduce along nx. When ReduceMode ==
 * kGlobalMode, use shared memory to reduce between threads.
 *
 * @template paraments
 * T: The type of data.
 * NX: The number of data continuously loaded by each thread.
 * NY: The number of data rows loaded by each thread, only NY = 1 was supported.
 * BlockSize: Identifies the current device thread index method. For GPU,
 * threadIdx.x is used as the thread index, and for xpu, core_id() is used as
 * the index. Currently only GPU was supported.
 * ReduceFunctor: Compute functor which has an operator() as following
 *     template <typename InT>
 *     struct ReduceFunctor {
 *       HOSTDEVICE OutT operator()(const InT& a, const InT& b) const {
 *         return ...;
 *       }
 *     };
 * ReduceMode: Reduce mode, can be kLocalMode, kGlobalMode.
 *
 * @param
 * out: The register pointer of out, the size is NX * NY.
 * in: The register pointer of in, the size is NX * NY.
 * reducer: Compute function which was declared like ReduceFunctor<InT>().
 * reduce_last_dim: if the last dim gets involved in reduction.
 */
template <typename T, int NX, int NY, int BlockSize, class ReduceFunctor,
          details::ReduceMode Mode>
__device__ __forceinline__ void Reduce(T* out, const T* in,
                                       ReduceFunctor reducer,
                                       bool reduce_last_dim) {
  int block_index = blockDim.y;

  if (Mode == details::ReduceMode::kGlobalMode) {
    bool block_reduce_y = (!reduce_last_dim) && (block_index > 1);
    // when reduce is not required for the last dim, and reduce num has been
    // split into multiple threads
    if (block_reduce_y) {
#pragma unroll
      for (int i = 0; i < NY * NX; i++) {  // reduce along blockdim.y
        out[i] = details::BlockYReduce<T, ReduceFunctor>(out[i], reducer);
      }
    }

    // when last dimension need to be reduced
    if (reduce_last_dim) {
#pragma unroll
      for (int i = 0; i < NY * NX; i++) {  // reduce along blockDim.x
        out[i] = details::BlockXReduce<T, ReduceFunctor>(out[i], reducer);
      }
    }
  } else {  // else  kLocalMode
#pragma unroll
    for (int i = 0; i < NY; ++i) {
#pragma unroll
      for (int j = 0; j < NX; ++j) {
        out[i] = reducer(out[i], in[i * NX + j]);
      }
    }
  }
}

}  // namespace kernel_primitives
}  // namespace operators
}  // namespace paddle