use of com.tencent.angel.ml.math2.storage.LongDoubleVectorStorage in project angel by Tencent.
the class MixedBinaryInZAExecutor method apply.
private static Vector apply(CompLongDoubleVector v1, LongDoubleVector v2, Binary op) {
LongDoubleVector[] parts = v1.getPartitions();
Storage[] resParts = StorageSwitch.applyComp(v1, v2, op);
if (v2.isSparse()) {
ObjectIterator<Long2DoubleMap.Entry> iter = v2.getStorage().entryIterator();
if (v1.size() > v2.size()) {
long subDim = (v1.getDim() + v1.getNumPartitions() - 1) / v1.getNumPartitions();
while (iter.hasNext()) {
Long2DoubleMap.Entry entry = iter.next();
long idx = entry.getLongKey();
int pidx = (int) (idx / subDim);
long subidx = idx % subDim;
if (parts[pidx].hasKey(subidx)) {
((LongDoubleVectorStorage) resParts[pidx]).set(subidx, op.apply(parts[pidx].get(subidx), entry.getDoubleValue()));
}
}
} else {
long base = 0;
for (int i = 0; i < parts.length; i++) {
LongDoubleVector part = parts[i];
LongDoubleVectorStorage resPart = (LongDoubleVectorStorage) resParts[i];
if (part.isDense()) {
double[] partValues = part.getStorage().getValues();
double[] resPartValues = resPart.getValues();
for (int j = 0; j < partValues.length; j++) {
if (v2.hasKey(j + base)) {
resPartValues[j] = op.apply(partValues[j], v2.get(j + base));
}
}
} else if (part.isSparse()) {
ObjectIterator<Long2DoubleMap.Entry> piter = part.getStorage().entryIterator();
while (piter.hasNext()) {
Long2DoubleMap.Entry entry = piter.next();
long idx = entry.getLongKey();
if (v2.hasKey(idx + base)) {
resPart.set(idx, op.apply(entry.getDoubleValue(), v2.get(idx + base)));
}
}
} else {
// sorted
if (op.isKeepStorage()) {
long[] partIndices = part.getStorage().getIndices();
double[] partValues = part.getStorage().getValues();
long[] resPartIndices = resPart.getIndices();
double[] resPartValues = resPart.getValues();
for (int j = 0; j < partIndices.length; j++) {
long idx = partIndices[j];
if (v2.hasKey(idx + base)) {
resPartIndices[j] = idx;
resPartValues[j] = op.apply(partValues[j], v2.get(idx + base));
}
}
} else {
long[] partIndices = part.getStorage().getIndices();
double[] partValues = part.getStorage().getValues();
for (int j = 0; j < partIndices.length; j++) {
long idx = partIndices[j];
if (v2.hasKey(idx + base)) {
resPart.set(idx, op.apply(partValues[j], v2.get(idx + base)));
}
}
}
}
base += part.getDim();
}
}
} else {
// sorted
if (v1.size() > v2.size()) {
long subDim = (v1.getDim() + v1.getNumPartitions() - 1) / v1.getNumPartitions();
long[] v2Indices = v2.getStorage().getIndices();
double[] v2Values = v2.getStorage().getValues();
for (int i = 0; i < v2Indices.length; i++) {
long idx = v2Indices[i];
int pidx = (int) (idx / subDim);
long subidx = idx % subDim;
if (parts[pidx].hasKey(subidx)) {
((LongDoubleVectorStorage) resParts[pidx]).set(subidx, op.apply(parts[pidx].get(subidx), v2Values[i]));
}
}
} else {
long base = 0;
for (int i = 0; i < parts.length; i++) {
LongDoubleVector part = parts[i];
LongDoubleVectorStorage resPart = (LongDoubleVectorStorage) resParts[i];
if (part.isDense()) {
double[] partValues = part.getStorage().getValues();
double[] resPartValues = resPart.getValues();
for (int j = 0; j < partValues.length; j++) {
if (v2.hasKey(j + base)) {
resPartValues[j] = op.apply(partValues[j], v2.get(j + base));
}
}
} else if (part.isSparse()) {
ObjectIterator<Long2DoubleMap.Entry> piter = part.getStorage().entryIterator();
while (piter.hasNext()) {
Long2DoubleMap.Entry entry = piter.next();
long idx = entry.getLongKey();
if (v2.hasKey(idx + base)) {
resPart.set(idx, op.apply(entry.getDoubleValue(), v2.get(idx + base)));
}
}
} else {
// sorted
if (op.isKeepStorage()) {
long[] partIndices = part.getStorage().getIndices();
double[] partValues = part.getStorage().getValues();
long[] resPartIndices = resPart.getIndices();
double[] resPartValues = resPart.getValues();
for (int j = 0; j < partIndices.length; j++) {
long idx = partIndices[j];
if (v2.hasKey(idx + base)) {
resPartIndices[j] = idx;
resPartValues[j] = op.apply(partValues[j], v2.get(idx + base));
}
}
} else {
long[] partIndices = part.getStorage().getIndices();
double[] partValues = part.getStorage().getValues();
for (int j = 0; j < partIndices.length; j++) {
long idx = partIndices[j];
if (v2.hasKey(idx + base)) {
resPart.set(idx, op.apply(partValues[j], v2.get(idx + base)));
}
}
}
}
base += part.getDim();
}
}
}
LongDoubleVector[] res = new LongDoubleVector[parts.length];
int i = 0;
for (LongDoubleVector part : parts) {
res[i] = new LongDoubleVector(part.getMatrixId(), part.getRowId(), part.getClock(), part.getDim(), (LongDoubleVectorStorage) resParts[i]);
i++;
}
v1.setPartitions(res);
return v1;
}
use of com.tencent.angel.ml.math2.storage.LongDoubleVectorStorage in project angel by Tencent.
the class MixedBinaryInNonZAExecutor method apply.
private static Vector apply(CompLongDoubleVector v1, LongDoubleVector v2, Binary op) {
LongDoubleVector[] parts = v1.getPartitions();
Storage[] resParts = StorageSwitch.applyComp(v1, v2, op);
if (v2.isSparse()) {
if (!op.isKeepStorage()) {
for (int i = 0; i < parts.length; i++) {
if (parts[i].getStorage() instanceof LongDoubleSortedVectorStorage) {
resParts[i] = new LongDoubleSparseVectorStorage(parts[i].getDim(), parts[i].getStorage().getIndices(), parts[i].getStorage().getValues());
}
}
}
long subDim = (v1.getDim() + v1.getNumPartitions() - 1) / v1.getNumPartitions();
ObjectIterator<Long2DoubleMap.Entry> iter = v2.getStorage().entryIterator();
while (iter.hasNext()) {
Long2DoubleMap.Entry entry = iter.next();
long gidx = entry.getLongKey();
int pidx = (int) (gidx / subDim);
long subidx = gidx % subDim;
((LongDoubleVectorStorage) resParts[pidx]).set(subidx, op.apply(parts[pidx].get(subidx), entry.getDoubleValue()));
}
} else {
// sorted
if (!op.isKeepStorage()) {
for (int i = 0; i < parts.length; i++) {
if (parts[i].getStorage() instanceof LongDoubleSortedVectorStorage) {
resParts[i] = new LongDoubleSparseVectorStorage(parts[i].getDim(), parts[i].getStorage().getIndices(), parts[i].getStorage().getValues());
}
}
}
long subDim = (v1.getDim() + v1.getNumPartitions() - 1) / v1.getNumPartitions();
long[] v2Indices = v2.getStorage().getIndices();
double[] v2Values = v2.getStorage().getValues();
for (int i = 0; i < v2Indices.length; i++) {
long gidx = v2Indices[i];
int pidx = (int) (gidx / subDim);
long subidx = gidx % subDim;
((LongDoubleVectorStorage) resParts[pidx]).set(subidx, op.apply(parts[pidx].get(subidx), v2Values[i]));
}
}
LongDoubleVector[] res = new LongDoubleVector[parts.length];
int i = 0;
for (LongDoubleVector part : parts) {
res[i] = new LongDoubleVector(part.getMatrixId(), part.getRowId(), part.getClock(), part.getDim(), (LongDoubleVectorStorage) resParts[i]);
i++;
}
v1.setPartitions(res);
return v1;
}
use of com.tencent.angel.ml.math2.storage.LongDoubleVectorStorage in project angel by Tencent.
the class MixedBinaryInAllExecutor method apply.
private static Vector apply(CompLongDoubleVector v1, LongLongVector v2, Binary op) {
LongDoubleVector[] parts = v1.getPartitions();
Storage[] resParts = StorageSwitch.applyComp(v1, v2, op);
if (!op.isKeepStorage()) {
for (int i = 0; i < parts.length; i++) {
if (parts[i].getStorage() instanceof LongDoubleSortedVectorStorage) {
resParts[i] = new LongDoubleSparseVectorStorage(parts[i].getDim(), parts[i].getStorage().getIndices(), parts[i].getStorage().getValues());
}
}
}
long subDim = (v1.getDim() + v1.getNumPartitions() - 1) / v1.getNumPartitions();
for (int i = 0; i < v1.getDim(); i++) {
int pidx = (int) (i / subDim);
long subidx = i % subDim;
if (v2.getStorage().hasKey(i)) {
((LongDoubleVectorStorage) resParts[pidx]).set(subidx, op.apply(parts[pidx].get(subidx), v2.get(i)));
} else {
((LongDoubleVectorStorage) resParts[pidx]).set(subidx, op.apply(parts[pidx].get(subidx), 0));
}
}
LongDoubleVector[] res = new LongDoubleVector[parts.length];
int i = 0;
for (LongDoubleVector part : parts) {
res[i] = new LongDoubleVector(part.getMatrixId(), part.getRowId(), part.getClock(), part.getDim(), (LongDoubleVectorStorage) resParts[i]);
i++;
}
v1.setPartitions(res);
return v1;
}
use of com.tencent.angel.ml.math2.storage.LongDoubleVectorStorage in project angel by Tencent.
the class LongDoubleVector method argmin.
public long argmin() {
LongDoubleVectorStorage idstorage = (LongDoubleVectorStorage) storage;
if (idstorage.size() == 0)
return -1;
double minval = Double.MAX_VALUE;
long minidx = -1;
if (idstorage.isSparse()) {
ObjectIterator<Long2DoubleMap.Entry> iter = idstorage.entryIterator();
while (iter.hasNext()) {
Long2DoubleMap.Entry entry = iter.next();
long idx = entry.getLongKey();
double val = entry.getDoubleValue();
if (val < minval) {
minval = val;
minidx = idx;
}
}
} else {
long[] indices = idstorage.getIndices();
double[] val = idstorage.getValues();
long size = idstorage.size();
for (int i = 0; i < size; i++) {
long idx = indices[i];
if (val[i] < minval) {
minval = val[i];
minidx = idx;
}
}
}
return minidx;
}
use of com.tencent.angel.ml.math2.storage.LongDoubleVectorStorage in project angel by Tencent.
the class LongDoubleVector method std.
public double std() {
LongDoubleVectorStorage dstorage = (LongDoubleVectorStorage) storage;
if (dstorage.size() == 0)
return 0;
double sumval = 0.0;
double sumval2 = 0.0;
if (dstorage.isSparse()) {
DoubleIterator iter = dstorage.valueIterator();
while (iter.hasNext()) {
double val = iter.nextDouble();
sumval += val;
sumval2 += val * val;
}
} else {
for (double val : dstorage.getValues()) {
sumval += val;
sumval2 += val * val;
}
}
sumval /= getDim();
sumval2 /= getDim();
return Math.sqrt(sumval2 - sumval * sumval);
}
Aggregations