/*******************************************************************************
* Copyright (c) 2015-2018 Skymind, Inc.
*
* This program and the accompanying materials are made available under the
* terms of the Apache License, Version 2.0 which is available at
* https://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.
*
* SPDX-License-Identifier: Apache-2.0
******************************************************************************/
//
// @author Yurii Shyrma, created on 21.09.2018
// @author raver119@gmail.com
//
#include <helpers/TAD.h>
#include<ops/declarable/helpers/ismax.h>
#include <helpers/ConstantTadHelper.h>
namespace nd4j {
namespace ops {
namespace helpers {
template <typename T>
static void ismax_(const NDArray* input, NDArray* output, const std::vector<int>& dimensions) {
if (input->isVector()) {
int dimensionsLength = dimensions.size();
int length = input->lengthOf();
if ((input->shapeOf())[dimensions[0]] == 1) {
for (int i = 0; i < length; i++)
output->p<T>(i, 1.f);
}
else {
int eleStride = shape::elementWiseStride(input->getShapeInfo());
if (eleStride == 1) {
int maxIdx = 0;
T currMax = input->e<T>(0);
if (length < ELEMENT_THRESHOLD) {
for (int i = 0; i < length; i++) {
if (currMax < input->e<T>(i)) {
currMax = input->e<T>(i);
maxIdx = i;
}
output->p<T>(i, 0.f);
}
}
else {
{
int maxIdxLocal = maxIdx;
T currMaxLocal = currMax;
for (int i = 0; i < length; i++) {
if (currMaxLocal < input->e<T>(i)) {
currMaxLocal = input->e<T>(i);
maxIdxLocal = i;
}
output->p<T>(i, 0.f);
}
PRAGMA_OMP_CRITICAL
{
if (currMax < currMaxLocal) {
currMax = currMaxLocal;
maxIdx = maxIdxLocal;
}
}
}
}
output->p<T>(maxIdx, 1.f);
}
else {
int maxIdx = 0;
T currMax = input->e<T>(0);
if (length < ELEMENT_THRESHOLD) {
for (int i = 0; i < length; i++) {
if (currMax < input->e<T>(i*eleStride)) {
currMax = input->e<T>(i*eleStride);
maxIdx = i;
}
output->p<T>(i, 0.f);
}
}
else {
{
int maxIdxLocal = maxIdx;
T currMaxLocal = currMax;
for (int i = 0; i < length; i++) {
if (currMaxLocal < input->e<T>(i*eleStride)) {
currMaxLocal = input->e<T>(i*eleStride);
maxIdxLocal = i;
}
output->p<T>(i, 0.f);
}
PRAGMA_OMP_CRITICAL
{
if (currMax < currMaxLocal) {
currMax = currMaxLocal;
maxIdx = maxIdxLocal;
}
}
}
}
output->p<T>(maxIdx, 1.f);
}
}
}
else {
int dimensionsLength = dimensions.size();
//int tads = tad.numTads;
//decompose in to several sub tads after
//moving all dimensions (in sorted order)
//to the back.
//permuted version of the input shape info for setting up the tad problem
auto tadPack = nd4j::ConstantTadHelper::getInstance()->tadForDimensions(input->getShapeInfo(), const_cast<int*>(dimensions.data()), dimensionsLength);
auto tadShapeShapeInfo = tadPack.primaryShapeInfo();
auto tadOffsets = tadPack.primaryOffsets();
int tadLength = shape::length(tadShapeShapeInfo);
int tads = tadPack.numberOfTads();
int tadsPerThread = tads / TAD_THRESHOLD;
int num_threads = nd4j::math::nd4j_max<int>(1, tadsPerThread);
num_threads = nd4j::math::nd4j_min<int>(num_threads, omp_get_max_threads());
auto tadEWS = shape::elementWiseStride(tadShapeShapeInfo);
auto zEWS = tadEWS;
int span = (tads / num_threads) + 8;
PRAGMA_OMP_PARALLEL_THREADS(num_threads)
{
int tid = omp_get_thread_num();
int start = span * tid;
int end = span * (tid + 1);
if (end > tads) end = tads;
for (int r = start; r < end; r++) {
if (tadEWS > 0 && zEWS > 0 && dimensionsLength == 1) {
T *rX = const_cast<NDArray*>(input)->bufferAsT<T>() + tadOffsets[r];
T *rZ = output->bufferAsT<T>() + tadOffsets[r];
T maxValue = rX[0];
int maxIdx = 0;
if (tadEWS == 1 && zEWS == 1) {
for (int i = 0; i < tadLength; i++) {
if (rX[i] > maxValue) {
maxIdx = i;
maxValue = rX[i];
}
}
PRAGMA_OMP_SIMD
for (int i = 0; i < tadLength; i++) {
rZ[i] = maxIdx == i ? (T) 1.0 : (T) 0.0;
}
}
else {
for (int i = 0; i < tadLength; i++) {
if (rX[i * tadEWS] > maxValue) {
maxIdx = i;
maxValue = rX[i * tadEWS];
}
}
PRAGMA_OMP_SIMD
for (int i = 0; i < tadLength; i++) {
rZ[i * zEWS] = maxIdx == i ? (T) 1.0 : (T) 0.0;
}
}
}
else {
int tadsPerThread = tads / TAD_THRESHOLD;
int num_threads = nd4j::math::nd4j_max<int>(1, tadsPerThread);
num_threads = nd4j::math::nd4j_min<int>(num_threads, omp_get_max_threads());
Nd4jLong offset = tadOffsets[r];
Nd4jLong shapeIter[MAX_RANK];
Nd4jLong coord[MAX_RANK];
int dim;
Nd4jLong xStridesIter[MAX_RANK];
Nd4jLong resultStridesIter[MAX_RANK];
Nd4jLong *xShape = shape::shapeOf(tadShapeShapeInfo);
Nd4jLong *xStride = shape::stride(tadShapeShapeInfo);
Nd4jLong *resultStride = shape::stride(tadShapeShapeInfo);
int rank = shape::rank(tadShapeShapeInfo);
T *xPointer = const_cast<NDArray*>(input)->bufferAsT<T>() + offset;
T *resultPointer = output->bufferAsT<T>() + offset;
T maxValue = xPointer[0];
T *maxCursor = resultPointer;
Nd4jPointer maxCursorLong = reinterpret_cast<Nd4jPointer>(maxCursor);
if (PrepareTwoRawArrayIter<T>(rank, xShape, xPointer, xStride, resultPointer, resultStride, &rank, shapeIter, &xPointer, xStridesIter, &resultPointer, resultStridesIter) >= 0) {
ND4J_RAW_ITER_START(dim, rank, coord, shapeIter);
{
if (maxValue < xPointer[0]) {
maxCursor = resultPointer;
maxCursorLong = reinterpret_cast<Nd4jPointer>(resultPointer);
maxValue = xPointer[0];
}
resultPointer[0] = 0.0;
}
ND4J_RAW_ITER_TWO_NEXT(dim, rank, coord, shapeIter, xPointer, xStridesIter, resultPointer, resultStridesIter);
maxCursor = reinterpret_cast<T*>(maxCursorLong);
maxCursor[0] = 1.0;
}
}
}
}
}
}
void ismax(nd4j::LaunchContext * context, const NDArray *input, NDArray *output, const std::vector<int>& dimensions) {
BUILD_SINGLE_SELECTOR(input->dataType(), ismax_, (input, output, dimensions), LIBND4J_TYPES);
}
BUILD_SINGLE_TEMPLATE(template void ismax_, (const NDArray *input, NDArray *output, const std::vector<int>& dimensions), LIBND4J_TYPES);
}
}
}