#include <EcalPedestalsGPU.h>

Classes
struct	Product

Public Member Functions
	EcalPedestalsGPU (EcalPedestals const &)

const Product &	getProduct (cudaStream_t) const

	~EcalPedestalsGPU ()=default

Static Public Member Functions
static std::string	name ()

Private Attributes
std::vector< float, cms::cuda::HostAllocator< float > >	mean_x12_

std::vector< float, cms::cuda::HostAllocator< float > >	mean_x1_

std::vector< float, cms::cuda::HostAllocator< float > >	mean_x6_

cms::cuda::ESProduct< Product >	product_

std::vector< float, cms::cuda::HostAllocator< float > >	rms_x12_

std::vector< float, cms::cuda::HostAllocator< float > >	rms_x1_

std::vector< float, cms::cuda::HostAllocator< float > >	rms_x6_

Detailed Description

Definition at line 13 of file EcalPedestalsGPU.h.

Constructor & Destructor Documentation

◆ EcalPedestalsGPU()

EcalPedestalsGPU::EcalPedestalsGPU ( EcalPedestals const & pedestals )

Definition at line 6 of file EcalPedestalsGPU.cc.

     : mean_x12_(pedestals.size()),
       rms_x12_(pedestals.size()),
       mean_x6_(pedestals.size()),
       rms_x6_(pedestals.size()),
       mean_x1_(pedestals.size()),
       rms_x1_(pedestals.size()) {
   // fill in eb
   auto const& barrelValues = pedestals.barrelItems();
   for (unsigned int i = 0; i < barrelValues.size(); i++) {
     mean_x12_[i] = barrelValues[i].mean_x12;
     rms_x12_[i] = barrelValues[i].rms_x12;
     mean_x6_[i] = barrelValues[i].mean_x6;
     rms_x6_[i] = barrelValues[i].rms_x6;
     mean_x1_[i] = barrelValues[i].mean_x1;
     rms_x1_[i] = barrelValues[i].rms_x1;
   }
  
   // fill in ee
   auto const& endcapValues = pedestals.endcapItems();
   auto const offset = barrelValues.size();
   for (unsigned int i = 0; i < endcapValues.size(); i++) {
     mean_x12_[offset + i] = endcapValues[i].mean_x12;
     rms_x12_[offset + i] = endcapValues[i].rms_x12;
     mean_x6_[offset + i] = endcapValues[i].mean_x6;
     rms_x6_[offset + i] = endcapValues[i].rms_x6;
     mean_x1_[offset + i] = endcapValues[i].mean_x1;
     rms_x1_[offset + i] = endcapValues[i].rms_x1;
   }
 }

References EcalCondObjectContainer< T >::barrelItems(), EcalCondObjectContainer< T >::endcapItems(), mps_fire::i, mean_x12_, mean_x1_, mean_x6_, hltrates_dqm_sourceclient-live_cfg::offset, rms_x12_, rms_x1_, and rms_x6_.

◆ ~EcalPedestalsGPU()

EcalPedestalsGPU::~EcalPedestalsGPU ( )

default

Member Function Documentation

◆ getProduct()

EcalPedestalsGPU::Product const & EcalPedestalsGPU::getProduct ( cudaStream_t cudaStream ) const

Definition at line 37 of file EcalPedestalsGPU.cc.

                                                                                          {
   auto const& product = product_.dataForCurrentDeviceAsync(
       cudaStream, [this](EcalPedestalsGPU::Product& product, cudaStream_t cudaStream) {
         // allocate
         product.mean_x12 = cms::cuda::make_device_unique<float[]>(mean_x12_.size(), cudaStream);
         product.mean_x6 = cms::cuda::make_device_unique<float[]>(mean_x6_.size(), cudaStream);
         product.mean_x1 = cms::cuda::make_device_unique<float[]>(mean_x1_.size(), cudaStream);
         product.rms_x12 = cms::cuda::make_device_unique<float[]>(rms_x12_.size(), cudaStream);
         product.rms_x6 = cms::cuda::make_device_unique<float[]>(rms_x6_.size(), cudaStream);
         product.rms_x1 = cms::cuda::make_device_unique<float[]>(rms_x1_.size(), cudaStream);
         // transfer
         cms::cuda::copyAsync(product.mean_x12, mean_x12_, cudaStream);
         cms::cuda::copyAsync(product.mean_x6, mean_x6_, cudaStream);
         cms::cuda::copyAsync(product.mean_x1, mean_x1_, cudaStream);
         cms::cuda::copyAsync(product.rms_x12, rms_x12_, cudaStream);
         cms::cuda::copyAsync(product.rms_x6, rms_x6_, cudaStream);
         cms::cuda::copyAsync(product.rms_x1, rms_x1_, cudaStream);
       });
  
   return product;
 }