diff --git a/RTCP/CNProc/src/CN_Processing.cc b/RTCP/CNProc/src/CN_Processing.cc
index 380c723d0192c1753ee00857c018215096cf8348..1896c7f1a9daa8ccc12e49743fd4fb076bbbc259 100644
--- a/RTCP/CNProc/src/CN_Processing.cc
+++ b/RTCP/CNProc/src/CN_Processing.cc
@@ -516,14 +516,17 @@ template <typename SAMPLE_TYPE> void CN_Processing<SAMPLE_TYPE>::formPencilBeams
computeTimer.stop();
}
-template <typename SAMPLE_TYPE> void CN_Processing<SAMPLE_TYPE>::calculateStokes()
+template <typename SAMPLE_TYPE> void CN_Processing<SAMPLE_TYPE>::calculateIncoherentStokes()
{
computeTimer.start();
- if( itsMode.isCoherent() ) {
- itsStokes->calculateCoherent(itsPencilBeamData,itsStokesData,itsPencilBeamFormer->nrCoordinates());
- } else {
- itsStokes->calculateIncoherent(itsFilteredData,itsStokesData,itsNrStations);
- }
+ itsStokes->calculateIncoherent(itsFilteredData,itsStokesData,itsNrStations);
+ computeTimer.stop();
+}
+
+template <typename SAMPLE_TYPE> void CN_Processing<SAMPLE_TYPE>::calculateCoherentStokes()
+{
+ computeTimer.start();
+ itsStokes->calculateCoherent(itsPencilBeamData,itsStokesData,itsPencilBeamFormer->nrCoordinates());
computeTimer.stop();
}
@@ -591,11 +594,13 @@ template <typename SAMPLE_TYPE> void CN_Processing<SAMPLE_TYPE>::process()
case CN_Mode::COHERENT_STOKES_I:
case CN_Mode::COHERENT_ALLSTOKES:
formPencilBeams();
- // fallthrough to incoherent modes
+ calculateCoherentStokes();
+ sendOutput( itsStokesData );
+ break;
case CN_Mode::INCOHERENT_STOKES_I:
case CN_Mode::INCOHERENT_ALLSTOKES:
- calculateStokes();
+ calculateIncoherentStokes();
sendOutput( itsStokesData );
break;
diff --git a/RTCP/CNProc/src/PencilBeams.cc b/RTCP/CNProc/src/PencilBeams.cc
index 6c5b0f2f34061fdd03e3160cac377bd018931731..4410fea5f809662ae17b3faaa77f5f6ff38e1b41 100644
--- a/RTCP/CNProc/src/PencilBeams.cc
+++ b/RTCP/CNProc/src/PencilBeams.cc
@@ -220,25 +220,43 @@ fcomplex PencilBeams::phaseShift( double frequency, double delay )
return makefcomplex( std::cos(phi), std::sin(phi) );
}
-void PencilBeams::computeComplexVoltages( MultiDimArray<fcomplex,4> &in, MultiDimArray<fcomplex,4> &out, std::vector<unsigned> stations )
+void PencilBeams::computeComplexVoltages( FilteredData *in, PencilBeamData *out, std::vector<unsigned> stations )
{
float factor = 1.0 / stations.size();
+ /* TODO: filter out stations with too much flagged data */
+
+ /* conservative flagging: flag output if any input was flagged */
+ for( unsigned beam = 0; beam < itsCoordinates.size(); beam++ ) {
+ out->flags[beam].reset();
+
+ for (unsigned stat = 0; stat < stations.size(); stat++ ) {
+ out->flags[beam] |= in->flags[stations[stat]];
+ }
+ }
+
for (unsigned ch = 0; ch < itsNrChannels; ch ++) {
double frequency = itsBaseFrequency + ch * itsChannelBandwidth;
for (unsigned time = 0; time < itsNrSamplesPerIntegration; time ++) {
for (unsigned pol = 0; pol < NR_POLARIZATIONS; pol ++) {
for( unsigned beam = 0; beam < itsCoordinates.size(); beam++ ) {
- fcomplex &dest = out[ch][beam][time][pol];
- fcomplex sample = makefcomplex( 0, 0 );
-
- for( unsigned stat = 0; stat < stations.size(); stat++ ) {
- // note: for beam #0 (central beam), the shift is 1
- fcomplex shift = phaseShift( frequency, itsDelays[stations[stat]][beam] );
- sample += in[ch][stations[stat]][time][pol] * shift;
+ fcomplex &dest = out->samples[ch][beam][time][pol];
+
+ if( !out->flags[beam].test(time) ) {
+ /* combine the stations for this beam */
+ fcomplex sample = makefcomplex( 0, 0 );
+
+ for( unsigned stat = 0; stat < stations.size(); stat++ ) {
+ // note: for beam #0 (central beam), the shift is 1
+ fcomplex shift = phaseShift( frequency, itsDelays[stations[stat]][beam] );
+ sample += in->samples[ch][stations[stat]][time][pol] * shift;
+ }
+
+ dest = sample * factor;
+ } else {
+ /* data is flagged */
+ dest = makefcomplex( 0, 0 );
}
-
- dest = sample * factor;
}
}
}
@@ -282,9 +300,7 @@ void PencilBeams::formPencilBeams( FilteredData *filteredData, PencilBeamData *p
stations.push_back( stat );
}
- std::clog << "Forming beams for stations " << stations << std::endl;
-
- computeComplexVoltages( filteredData->samples, pencilBeamData->samples, stations );
+ computeComplexVoltages( filteredData, pencilBeamData, stations );
}
} // namespace RTCP
diff --git a/RTCP/CNProc/src/PencilBeams.h b/RTCP/CNProc/src/PencilBeams.h
index 0f1a295a4c899eeef12f5144ec90e655ef68ce06..78b680091922657ca037d0dfbe286130cf0f9701 100644
--- a/RTCP/CNProc/src/PencilBeams.h
+++ b/RTCP/CNProc/src/PencilBeams.h
@@ -154,7 +154,7 @@ class PencilBeams
void calculateDelays( unsigned stat, const PencilCoord3D &beamDir );
void calculateAllDelays( FilteredData *filteredData );
- void computeComplexVoltages( MultiDimArray<fcomplex,4> &in, MultiDimArray<fcomplex,4> &out, std::vector<unsigned> stations );
+ void computeComplexVoltages( FilteredData *filteredData, PencilBeamData *pencilBeamData, std::vector<unsigned> stations );
std::vector<PencilCoord3D> itsCoordinates;
unsigned itsNrStations;
diff --git a/RTCP/CNProc/src/Stokes.cc b/RTCP/CNProc/src/Stokes.cc
index f91153dd89c45b2cdfaf6a61027015a6cb6c5d76..2e0ee08a842946d16150a43a26b93f4678074965 100644
--- a/RTCP/CNProc/src/Stokes.cc
+++ b/RTCP/CNProc/src/Stokes.cc
@@ -18,34 +18,55 @@ Stokes::Stokes( CN_Mode &mode, unsigned nrChannels, unsigned nrSamplesPerIntegra
void Stokes::calculateCoherent( PencilBeamData *pencilBeamData, StokesData *stokesData, unsigned nrBeams )
{
- computeStokes( pencilBeamData->samples, stokesData->samples, nrBeams );
+ computeStokes( pencilBeamData->samples, pencilBeamData->flags, stokesData, nrBeams );
}
void Stokes::calculateIncoherent( FilteredData *filteredData, StokesData *stokesData, unsigned nrStations )
{
- computeStokes( filteredData->samples, stokesData->samples, nrStations );
+ computeStokes( filteredData->samples, filteredData->flags, stokesData, nrStations );
}
static float sqr( float x ) { return x * x; }
-void Stokes::computeStokes( MultiDimArray<fcomplex,4> &in, MultiDimArray<float,4> &out, unsigned nrBeams )
+void Stokes::computeStokes( MultiDimArray<fcomplex,4> &in, SparseSet<unsigned> *inflags, StokesData *out, unsigned nrBeams )
{
unsigned &integrationSteps = itsNrSamplesPerStokesIntegration;
bool allStokes = itsNrStokes == 4;
bool coherent = itsIsCoherent;
+ unsigned nrOutputs = coherent ? nrBeams : 1;
- std::clog << "Calculating " << itsNrStokes << " Stokes for " << nrBeams << " beam(s)." << std::endl;
+ for(unsigned beam = 0; beam < nrOutputs; beam++) {
+ out->flags[beam].reset();
+ }
+
+ /* conservative flagging; flag output if any of the inputs are flagged */
+ /* TODO: fix for integrationsteps > 1 */
+ for(unsigned beam = 0; beam < nrBeams; beam++) {
+ out->flags[coherent ? beam : 0] |= inflags[beam];
+ }
+
+ for(unsigned beam = 0; beam < nrOutputs; beam++) {
+ out->flags[beam] /= integrationSteps;
+ }
for (unsigned ch = 0; ch < itsNrChannels; ch ++) {
for (unsigned time = 0; time < itsNrSamplesPerIntegration; time += integrationSteps ) {
float stokesI = 0, stokesQ = 0, stokesU = 0, stokesV = 0;
+ float nrValidSamples = 0;
+ bool flagged = false;
for( unsigned beam = 0; beam < nrBeams; beam++ ) {
if( coherent ) {
stokesI = stokesQ = stokesU = stokesV = 0;
+ nrValidSamples = 0;
+ flagged = false;
}
for( unsigned fractime = 0; fractime < integrationSteps; fractime++ ) {
+ if( inflags[beam].test( time+fractime ) ) {
+ continue;
+ }
+
// assert: two polarizations
fcomplex &polX = in[ch][beam][time+fractime][0];
@@ -59,34 +80,42 @@ void Stokes::computeStokes( MultiDimArray<fcomplex,4> &in, MultiDimArray<float,4
stokesU += real(polX * conj(polY));
stokesV += imag(polX * conj(polY));
}
+ nrValidSamples++;
}
- #define dest out[ch][beam][time / integrationSteps]
if( coherent ) {
- dest[0] = stokesI;
+ if( flagged ) {
+ /* hack: if no valid samples, insert zeroes */
+ if( nrValidSamples == 0 ) { nrValidSamples = 1; }
+ }
+ #define dest out->samples[ch][beam][time / integrationSteps]
+ dest[0] = stokesI / nrValidSamples;
if( allStokes ) {
- dest[1] = stokesQ;
- dest[2] = stokesU;
- dest[3] = stokesV;
+ dest[1] = stokesQ / nrValidSamples;
+ dest[2] = stokesU / nrValidSamples;
+ dest[3] = stokesV / nrValidSamples;
}
+ #undef dest
}
- #undef dest
}
if( !coherent ) {
- #define dest out[ch][0][time / integrationSteps]
- dest[0] = stokesI;
+ if( flagged ) {
+ /* hack: if no valid samples, insert zeroes */
+ if( nrValidSamples == 0 ) { nrValidSamples = 1; }
+ }
+
+ #define dest out->samples[ch][0][time / integrationSteps]
+ dest[0] = stokesI / nrValidSamples;
if( allStokes ) {
- dest[1] = stokesQ;
- dest[2] = stokesU;
- dest[3] = stokesV;
+ dest[1] = stokesQ / nrValidSamples;
+ dest[2] = stokesU / nrValidSamples;
+ dest[3] = stokesV / nrValidSamples;
}
#undef dest
}
}
}
-
- std::clog << "Done calculating Stokes" << std::endl;
}
} // namespace RTCP
diff --git a/RTCP/CNProc/src/Stokes.h b/RTCP/CNProc/src/Stokes.h
index 893982a46a41d838ef8f250eb846bce968fc751e..ff8b6000d1f1eb2673fdbedaa510fe2e1a230289 100644
--- a/RTCP/CNProc/src/Stokes.h
+++ b/RTCP/CNProc/src/Stokes.h
@@ -26,7 +26,7 @@ class Stokes
unsigned itsNrStokes;
bool itsIsCoherent;
- void computeStokes( MultiDimArray<fcomplex,4> &in, MultiDimArray<float,4> &out, unsigned nrBeams );
+ void computeStokes( MultiDimArray<fcomplex,4> &in, SparseSet<unsigned> *inflags, StokesData *out, unsigned nrBeams );
};
} // namespace RTCP
diff --git a/RTCP/Interface/include/Interface/FilteredData.h b/RTCP/Interface/include/Interface/FilteredData.h
index cbd4db9d470422e5c99d6a4552902d62d57eafce..5fa088b55930c25185f7e5910779243bcded7d44 100644
--- a/RTCP/Interface/include/Interface/FilteredData.h
+++ b/RTCP/Interface/include/Interface/FilteredData.h
@@ -13,27 +13,20 @@
namespace LOFAR {
namespace RTCP {
-class FilteredData: public StreamableData
+class FilteredData: public SampleData<fcomplex,4>
{
public:
+ typedef SampleData<fcomplex,4> SuperType;
+
FilteredData(unsigned nrStations, unsigned nrChannels, unsigned nrSamplesPerIntegration, Allocator &allocator = heapAllocator);
- ~FilteredData();
static size_t requiredSize(unsigned nrStations, unsigned nrChannels, unsigned nrSamplesPerIntegration);
- // The "| 2" significantly improves transpose speeds for particular
- // numbers of stations due to cache conflict effects. The extra memory
- // is not used.
- MultiDimArray<fcomplex, 4> samples; //[itsNrChannels][itsNrStations][itsNrSamplesPerIntegration | 2][NR_POLARIZATIONS] CACHE_ALIGNED
- SparseSet<unsigned> *flags; //[itsNrStations]
Vector<SubbandMetaData> metaData; //[itsNrStations]
protected:
virtual void readData( Stream* );
virtual void writeData( Stream* );
-
- private:
- void checkEndianness();
};
@@ -46,45 +39,25 @@ inline size_t FilteredData::requiredSize(unsigned nrStations, unsigned nrChannel
inline FilteredData::FilteredData(unsigned nrStations, unsigned nrChannels, unsigned nrSamplesPerIntegration, Allocator &allocator)
:
- StreamableData(false),
- samples(boost::extents[nrChannels][nrStations][nrSamplesPerIntegration | 2][NR_POLARIZATIONS], 32, allocator),
- flags(new SparseSet<unsigned>[nrStations]),
+ // The "| 2" significantly improves transpose speeds for particular
+ // numbers of stations due to cache conflict effects. The extra memory
+ // is not used.
+ SuperType::SampleData(false,boost::extents[nrChannels][nrStations][nrSamplesPerIntegration | 2][NR_POLARIZATIONS], nrStations, allocator),
metaData(nrStations, 32, allocator)
{
}
-
-inline FilteredData::~FilteredData()
-{
- delete [] flags;
-}
-
-
-inline void FilteredData::checkEndianness()
-{
-#if !defined WORDS_BIGENDIAN
- dataConvert(LittleEndian, samples.origin(), samples.num_elements());
-#endif
-}
-
-
inline void FilteredData::readData(Stream *str)
{
- str->read(samples.origin(), samples.num_elements() * sizeof(fcomplex));
str->read(&metaData[0], metaData.size() * sizeof(SubbandMetaData));
-
- checkEndianness();
+ SuperType::readData(str);
}
inline void FilteredData::writeData(Stream *str)
{
-#if !defined WORDS_BIGENDIAN
- THROW(AssertError, "not implemented: think about endianness");
-#endif
-
- str->write(samples.origin(), samples.num_elements() * sizeof(fcomplex));
str->write(&metaData[0], metaData.size() * sizeof(SubbandMetaData));
+ SuperType::writeData(str);
}
diff --git a/RTCP/Interface/include/Interface/PencilBeamData.h b/RTCP/Interface/include/Interface/PencilBeamData.h
index 9a2f7eff3eb96ca2385119cdeef6e5bcf271f338..abf9e55bfbc57509e799ce6c0070704947c7c3fb 100644
--- a/RTCP/Interface/include/Interface/PencilBeamData.h
+++ b/RTCP/Interface/include/Interface/PencilBeamData.h
@@ -13,64 +13,29 @@
namespace LOFAR {
namespace RTCP {
-class PencilBeamData: public StreamableData
+class PencilBeamData: public SampleData<fcomplex,4>
{
public:
- PencilBeamData(unsigned nrStations, unsigned nrChannels, unsigned nrSamplesPerIntegration, Allocator &allocator = heapAllocator);
+ typedef SampleData<fcomplex,4> SuperType;
+ PencilBeamData(unsigned nrStations, unsigned nrChannels, unsigned nrSamplesPerIntegration, Allocator &allocator = heapAllocator);
static size_t requiredSize(unsigned nrCoordinates, unsigned nrChannels, unsigned nrSamplesPerIntegration);
-
- // The "| 2" significantly improves transpose speeds for particular
- // numbers of stations due to cache conflict effects. The extra memory
- // is not used.
- MultiDimArray<fcomplex, 4> samples; //[itsNrChannels][itsNrCoordinates][itsNrSamplesPerIntegration | 2][NR_POLARIZATIONS] CACHE_ALIGNED
-
- protected:
- virtual void readData( Stream* );
- virtual void writeData( Stream* );
-
- private:
- void checkEndianness();
};
-
inline size_t PencilBeamData::requiredSize(unsigned nrCoordinates, unsigned nrChannels, unsigned nrSamplesPerIntegration)
{
return align(sizeof(fcomplex) * nrChannels * nrCoordinates * (nrSamplesPerIntegration | 2) * NR_POLARIZATIONS, 32);
}
inline PencilBeamData::PencilBeamData(unsigned nrCoordinates, unsigned nrChannels, unsigned nrSamplesPerIntegration, Allocator &allocator)
+ // The "| 2" significantly improves transpose speeds for particular
+ // numbers of stations due to cache conflict effects. The extra memory
+ // is not used.
:
- StreamableData(false),
- samples(boost::extents[nrChannels][nrCoordinates][nrSamplesPerIntegration | 2][NR_POLARIZATIONS], 32, allocator)
+ SuperType::SampleData(false, boost::extents[nrChannels][nrCoordinates][nrSamplesPerIntegration | 2][NR_POLARIZATIONS], nrCoordinates, allocator )
{
}
-inline void PencilBeamData::checkEndianness()
-{
-#if !defined WORDS_BIGENDIAN
- dataConvert(LittleEndian, samples.origin(), samples.num_elements());
-#endif
-}
-
-
-inline void PencilBeamData::readData(Stream *str)
-{
- str->read(samples.origin(), samples.num_elements() * sizeof(fcomplex));
-
- checkEndianness();
-}
-
-
-inline void PencilBeamData::writeData(Stream *str)
-{
-#if !defined WORDS_BIGENDIAN
- THROW(AssertError, "not implemented: think about endianness");
-#endif
-
- str->write(samples.origin(), samples.num_elements() * sizeof(fcomplex));
-}
-
} // namespace RTCP
} // namespace LOFAR
diff --git a/RTCP/Interface/include/Interface/StokesData.h b/RTCP/Interface/include/Interface/StokesData.h
index dcd73a9c211a40cde73dfbbaebac46650435dce7..e0d8d411db8ba4dbf3d93f024f2144f92108cf97 100644
--- a/RTCP/Interface/include/Interface/StokesData.h
+++ b/RTCP/Interface/include/Interface/StokesData.h
@@ -14,26 +14,14 @@
namespace LOFAR {
namespace RTCP {
-class StokesData: public StreamableData
+class StokesData: public SampleData<float,4>
{
public:
+ typedef SampleData<float,4> SuperType;
+
StokesData(CN_Mode &mode, unsigned nrPencilBeams, unsigned nrChannels, unsigned nrSamplesPerIntegration, unsigned nrSamplesPerStokesIntegration, Allocator &allocator = heapAllocator);
static size_t requiredSize(CN_Mode &mode, unsigned nrPencilBeams, unsigned nrChannels, unsigned nrSamplesPerIntegration, unsigned nrSamplesPerStokesIntegration);
-
- // The "| 2" significantly improves transpose speeds for particular
- // numbers of stations due to cache conflict effects. The extra memory
- // is not used.
- MultiDimArray<float, 4> samples; //[itsNrChannels][nrBeams()][itsNrSamplesPerIntegration | 2][nrStokes()] CACHE_ALIGNED
-
- protected:
- virtual void readData( Stream* );
- virtual void writeData( Stream* );
-
- private:
- bool itsHaveWarnedLittleEndian;
-
- void checkEndianness();
};
inline size_t StokesData::requiredSize(CN_Mode &mode, unsigned nrPencilBeams, unsigned nrChannels, unsigned nrSamplesPerIntegration, unsigned nrSamplesPerStokesIntegration)
@@ -43,43 +31,13 @@ inline size_t StokesData::requiredSize(CN_Mode &mode, unsigned nrPencilBeams, un
inline StokesData::StokesData(CN_Mode &mode, unsigned nrPencilBeams, unsigned nrChannels, unsigned nrSamplesPerIntegration, unsigned nrSamplesPerStokesIntegration, Allocator &allocator)
:
- StreamableData(false),
- samples(boost::extents[nrChannels][mode.isCoherent() ? nrPencilBeams : 1][(nrSamplesPerIntegration/nrSamplesPerStokesIntegration) | 2][mode.nrStokes()], 32, allocator),
- itsHaveWarnedLittleEndian(false)
-{
-}
-
-inline void StokesData::checkEndianness()
-{
-#if !defined WORDS_BIGENDIAN
- dataConvert(LittleEndian, samples.origin(), samples.num_elements());
-#endif
-}
-
-
-inline void StokesData::readData(Stream *str)
+ // The "| 2" significantly improves transpose speeds for particular
+ // numbers of stations due to cache conflict effects. The extra memory
+ // is not used.
+ SuperType::SampleData(false, boost::extents[nrChannels][mode.isCoherent() ? nrPencilBeams : 1][(nrSamplesPerIntegration/nrSamplesPerStokesIntegration) | 2][mode.nrStokes()], mode.isCoherent() ? nrPencilBeams : 1, allocator)
{
- str->read(samples.origin(), samples.num_elements() * sizeof(float));
-
- checkEndianness();
}
-
-inline void StokesData::writeData(Stream *str)
-{
-#if !defined WORDS_BIGENDIAN
- if( !itsHaveWarnedLittleEndian ) {
- itsHaveWarnedLittleEndian = true;
-
- std::clog << "Warning: writing data in little endian." << std::endl;
- }
- //THROW(AssertError, "not implemented: think about endianness");
-#endif
-
- str->write(samples.origin(), samples.num_elements() * sizeof(float));
-}
-
-
} // namespace RTCP
} // namespace LOFAR
diff --git a/RTCP/Interface/include/Interface/StreamableData.h b/RTCP/Interface/include/Interface/StreamableData.h
index 56c29fae5926a026772c85a25652a463d28def93..e910fcfd8d78fe6e4bbc6e84841cc0ca15a369d6 100644
--- a/RTCP/Interface/include/Interface/StreamableData.h
+++ b/RTCP/Interface/include/Interface/StreamableData.h
@@ -4,6 +4,8 @@
#include <Stream/Stream.h>
#include <Common/LofarTypes.h>
#include <Interface/Parset.h>
+#include <Interface/MultiDimArray.h>
+#include <Interface/SparseSet.h>
#include <Interface/Allocator.h>
#include <Common/DataConvert.h>
@@ -54,6 +56,27 @@ class StreamableData {
virtual void writeData(Stream*) = 0;
};
+// A typical data set contains a MultiDimArray of tuples and a set of flags.
+template <typename T, unsigned DIM> class SampleData: public StreamableData {
+ public:
+ typedef typename MultiDimArray<T,DIM>::ExtentList ExtentList;
+
+ SampleData( bool isIntegratable, const ExtentList &extents, unsigned nrFlags, Allocator &allocator = heapAllocator );
+ virtual ~SampleData();
+
+ MultiDimArray<T,DIM> samples;
+ SparseSet<unsigned> *flags;
+
+ protected:
+ virtual void checkEndianness();
+
+ virtual void readData(Stream*);
+ virtual void writeData(Stream*);
+
+ private:
+ bool itsHaveWarnedLittleEndian;
+};
+
inline void StreamableData::read( Stream *str, bool withSequenceNumber )
{
if( withSequenceNumber ) {
@@ -83,6 +106,49 @@ inline void StreamableData::write( Stream *str, bool withSequenceNumber )
writeData( str );
}
+template <typename T, unsigned DIM> inline SampleData<T,DIM>::SampleData( bool isIntegratable, const ExtentList &extents, unsigned nrFlags, Allocator &allocator ):
+ StreamableData( isIntegratable ),
+ samples( extents, 32, allocator ),
+ flags( new SparseSet<unsigned>[nrFlags] ),
+ itsHaveWarnedLittleEndian( false )
+{
+}
+
+template <typename T, unsigned DIM> inline SampleData<T,DIM>::~SampleData()
+{
+ delete [] flags;
+}
+
+
+template <typename T, unsigned DIM> inline void SampleData<T,DIM>::checkEndianness()
+{
+#if !defined WORDS_BIGENDIAN
+ dataConvert(LittleEndian, samples.origin(), samples.num_elements());
+#endif
+}
+
+template <typename T, unsigned DIM> inline void SampleData<T,DIM>::readData( Stream *str )
+{
+ str->read(samples.origin(), samples.num_elements() * sizeof (T) );
+
+ checkEndianness();
+}
+
+template <typename T, unsigned DIM> inline void SampleData<T,DIM>::writeData( Stream *str )
+{
+#if !defined WORDS_BIGENDIAN
+ if( !itsHaveWarnedLittleEndian ) {
+ itsHaveWarnedLittleEndian = true;
+
+ std::clog << "Warning: writing data in little endian." << std::endl;
+ }
+ //THROW(AssertError, "not implemented: think about endianness");
+#endif
+
+ str->write(samples.origin(), samples.num_elements() * sizeof (T) );
+}
+
+
} // namespace RTCP
} // namespace LOFAR