| DCMF::LwarxStwcxProcAtomic::__attribute__ | |
| DCMF::Device::WorkQueue::_BaseWorkQueue | Defines the interfaces and operations for producer/consumer work queues |
| _BGP_IP_Addr_t | |
| _BGP_Personality_DDR_t | |
| _BGP_Personality_Ethernet_t | |
| _BGP_Personality_Kernel_t | |
| _BGP_Personality_Networks_t | |
| _BGP_PersonalityTreeInfo_t | |
| _BGP_UCI_BulkPowerSupply_t | |
| _BGP_UCI_ClockCard_t | |
| _BGP_UCI_ComputeCard_t | |
| _BGP_UCI_DDRChip_t | |
| _BGP_UCI_ENetConnector_t | |
| _BGP_UCI_Fan_t | |
| _BGP_UCI_FanAssembly_t | |
| _BGP_UCI_IOCard_t | |
| _BGP_UCI_LinkCard_t | |
| _BGP_UCI_LinkChip_t | |
| _BGP_UCI_LinkPort_t | |
| _BGP_UCI_Midplane_t | |
| _BGP_UCI_NodeCard_t | |
| _BGP_UCI_PowerCable_t | |
| _BGP_UCI_PowerModule_t | |
| _BGP_UCI_Rack_t | |
| _BGP_UCI_ServiceCard_t | |
| _BGPTreePacketSoftHeader_tag | |
| _cheader_data | |
| _Kernel_Coordinates | Converts all ranks into a XYZT Coordinate |
| CCMI::Schedule::OneColorRectangle::_me_t | Structure for coordinates needed while creating the schedule |
| CCMI::Executor::AllreduceState::_phase_state | |
| DCMF::Protocol::MultiSend::_rootStruct | |
| DCMF::Collectives::Allreduce::RectangleRingDput::Composite::_se | |
| DCMF::Collectives::A2AProtocol | |
| DCMF::Protocol::Send::CDI::ack_info_t< T_DmaMemregion > | Acknowledgement information returned to end the send operation |
| DCMF::Protocol::Send::AckPacketInfo | |
| DCMF::Collectives::Adaptor | |
| DCMF::Allocator | |
| CCMI::Executor::Allreduce | Allreduce (and Reduce) basic general executor |
| CCMI::Executor::AllreduceBase | |
| CCMI::Executor::AllreduceState | Allreduce (and Reduce) persistent state data This class allocates storage for receive buffers and schedule data |
| DCMF::Collectives::AlltoallFactory | |
| DCMF::Collectives::Allreduce::AsyncComposite | The Async Composite for the Allreduce kernel executor |
| DCMF::Collectives::Broadcast::AsyncCompositeFactoryT< T, afn > | Base factory class for broadcast factory implementations |
| DCMF::Collectives::Allreduce::AsyncCompositeT< SCHEDULE, EXECUTOR > | Allreduce protocol composite |
| DCMF::Collectives::Broadcast::AsyncCompositeT< T, sfn > | Asyc Broadcast Composite. It is single color right now |
| DCMF::Collectives::Allreduce::AsyncFactory | Base class for allreduce factory implementations |
| DCMF::Collectives::Allreduce::AsyncFactoryT< CONNMGR, COMPOSITE > | Factory template class for CompositeT |
| DCMF::Atomic | Base Class for Atomic (counters) |
| DCMF::atomic_factory_t | Structure used to pass implementation parameters |
| DCMF::AtomicFactory | |
| CCMI::Schedule::axis_rect | Structure to store rectangle by axii_t |
| DCMF::Barrier | Base Class for Barrier |
| CCMI::Executor::Barrier | |
| DCMF::Collectives::Barrier::BarrierFactory | Barrier Factory Base class |
| DCMF::Collectives::Barrier::BarrierFactoryT< T > | Barrier Factory Base class |
| DCMF::Collectives::Barrier::BarrierR< R > | Barrier for rectangle schedules The R argument chooses a rectangle schedule |
| DCMF::Collectives::Barrier::BarrierT< S, afn > | Binomial barrier |
| DCMF::Collectives::Allreduce::BaseComposite | |
| DCMF::ChannelSys::BaseCount | |
| DCMF::BaseDevice | Base Class for Devices |
| DCMF::DMA::BcastMemFifoDesc | |
| DCMF::DMA::BcastPutDesc | |
| DCMF::Collectives::Broadcast::BcastQueueElem | |
| BGP_UPC_Read_Counter_Config_Struct | |
| BGP_UPC_Read_Counters_Struct | |
| DCMF::BGPLockManager | |
| DCMF::BGPMemregion | |
| DCMF::BGPMessager | Messager class for the BGP platform |
| CCMI::Schedule::BinomialTreeSchedule | |
| CCMI::Executor::Broadcast | |
| DCMF::Collectives::Broadcast::BroadcastFactory | Base factory class for synchronous broadcast factory implementations |
| CCMI_Callback_t | |
| CCMI::MultiSend::CCMI_Multicast_t | |
| CCMI::MultiSend::CCMI_MulticastRecv_t | |
| CCMIQuad | |
| DCMF::DMA::ChanDevice | This is an access class for the array of DMA::Device objects |
| DCMF::DMA::ChanInfo | |
| DCMF::ChannelSys::Channel | The ChannelSys::Channel object |
| DCMF::ChannelMgr | This object provides Channel and Comm-Thread management |
| DCMF::DMA::Collective< TDesc > | |
| CollTblEntry_t | |
| CCMI::ConnectionManager::ColorConnMgr | |
| CCMI::ConnectionManager::ColorGeometryConnMgr | |
| DCMF::CommThread | This object provides support for creating comm thread objects |
| CCMI::Executor::Composite | |
| DCMF::Collectives::Allreduce::Composite | The Composite for the Allreduce (and reduce) kernel executor |
| DCMF::Collectives::Allreduce::RectangleRingDput::Composite | Rectangle allreduce protocol |
| DCMF::Collectives::Allreduce::CompositeT< SCHEDULE, EXECUTOR > | Allreduce protocol composite |
| DCMF::Collectives::Allreduce::Tree::CompositeT< SCHEDULE, EXECUTOR > | Tree allreduce protocol |
| DCMF::Collectives::ConfigFlags | Configuration flags/options for creating the factory |
| CCMI::ConnectionManager::ConnectionManager | A class to manage multisend connections for collective protocols |
| DCMF::DMA::Control | Control |
| DCMF::Protocol::Control::ControlDMAFactory | Point-to-point DMA control factory |
| DCMF::Protocol::Control::ControlFactory | Base class for generic point-to-point control factory implementations |
| DCMF::Protocol::Control::ControlOverSendFactory | Point-to-point control factory |
| DCMF::DMA::CounterAttrs_t | Counter Attributes |
| DCMF::DMA::CounterClient | DMA Counter Client Class |
| DCMF::DMA::Device::CounterClient | |
| DCMF::DMA::CounterGroup | DMA Counter Group Class |
| DCMF::DMA::CounterList | DMA Counter List Class |
| DCMF::DMA::CounterQuantum_t | Counter Quantum Information |
| DCMF_Allreduce_Configuration_t | Allreduce configuration |
| DCMF_Alltoallv_Configuration_t | Alltoallv configuration |
| DCMF_Barrier_Configuration_t | Barrier configuration |
| DCMF_Broadcast_Configuration_t | Broadcast configuration |
| DCMF_Callback_t | Completion callback information descriptor |
| DCMF_Configure_t | The object used to pass around API Configuration info |
| DCMF_Control_Configuration_t | Control Protocol Configuration Information |
| DCMF_Error_t | DCMF Error callback results |
| DCMF_Get_Configuration_t | Point-to-point get configuration |
| DCMF_GlobalAllreduce_Configuration_t | Global Allreduce Protocol Configuration Information |
| DCMF_GlobalBarrier_Configuration_t | Global Barrier Protocol Configuration Information |
| DCMF_GlobalBcast_Configuration_t | Global Broadcast Protocol Configuration Information |
| DCMF_Hardware_t | Hardware information data type |
| DCMF_Iovec_t | Internal structure for handling vectored (non-config) transfers |
| DCMF_Manytomany_Configuration_t | Many to many configuration |
| DCMF_Messager_advance_options | |
| DCMF_Multicast_Configuration_t | Multicast configuration. The connection ids here go from 0 - (nconnections-1) |
| DCMF_Multicast_t | A structure to pass parameters for the multisend multicast operation |
| DCMF_MulticastRecv_t | A structure to pass parameters for the multisend multicast post recv operation |
| DCMF_NetworkCoord_t | A structure to describe a network coordinate |
| DCMF_Put_Configuration_t | 1-sided point-to-point put configuration |
| DCMF_Reduce_Configuration_t | Reduce configuration |
| DCMF_Send_Configuration_t | Point-to-point send configuration |
| DCQuad | |
| DCR_GlobalInt_t | Global Interupt DCR. Note that userEnables returns the uptree state in diagnostic read mode |
| DCR_TreeClassRoute_t | Tree class route descriptor register DCR for classes 2 and 3 |
| DCR_TreeInjectionExceptionEnable_t | Tree Injection Exception Enable Register |
| DCR_TreeInjectionExceptionFlag_t | Tree Injection Exception Flag Register |
| DCR_TreeReceptionExceptionEnable_t | Tree Reception Register DCR |
| DCR_TreeReceptionExceptoinFlag_t | Tree Reception Register DCR |
| DCMF::Protocol::Send::DefaultSendFactory | |
| DCMF::DMA::DetermGetDesc | |
| DCMF::DMA::DetermMemFifoDesc | |
| DCMF::DMA::Device | |
| DCMF::Queueing::GI::Device | A Device implmentation of a Queuesystem This class implements a queue system. Currently the Global interrupt device has a single queue. The user posts to the queue and the global interrupts are called in order until all of them have been executed |
| DCMF::Queueing::Lockbox::Device | A Device implmentation of a Queuesystem This class implements a queue system. Currently the Global interrupt device has a single queue. The user posts to the queue and the interprocess communications are called in order until all of them have been executed |
| DCMF::Queueing::Tree::Device | |
| DCMF::CDI::DMA::Device< T, M > | Common Device Interface for all dma devices |
| DCMF::CDI::Packet::Device< T, M > | Common Device Interface for all packet devices |
| DCMF::Queueing::Generic::Device | A Device implmentation of a Queuesystem This class implements a queue system. The user posts to the queue and the interprocess communications are called in order until all of them have been executed |
| DCMF::DMA::device_channel | |
| DCMF::DMA::Direct< TDesc > | The direct template sends data to a Memory FIFO. The template could be used to implement an Eager protocol. The template however assumes that the receive object has already been created when its first packet arrives. For MPI, a tag-matching header packet should be sent before this protocol by the method. The direct protocol has also been designed for copy-bit rectangular and software tree based broadcasts for short messages |
| DCMF::DMA::DirectRecv | |
| DCMF::CDI::dispatch_t | Dispatch array structure |
| DMA_AddressingAppSegment_t | __INLINE__ definition |
| DMA_AddressingAppSegmentSet_t | Application Segment Set |
| DMA_AddressingGlobalAppSegmentSetInfo_t | Global Application Segment Set Info |
| DMA_AddressingInitSharedMemoryHeader | Shared Memory Header |
| DMA_AddressingLocalAppSegmentSetInfo_t | Local Application Segment Set Info |
| DCMF::CDI::dma_chunk_info_t | |
| DMA_Counter_t | Software DMA Counter Structure |
| DMA_CounterAppSegment_t | Counter Application Segment |
| DMA_CounterGroup_t | DMA Counter Group Structure |
| DMA_CounterHw_t | Hardware DMA Counter |
| DMA_CounterStatus_t | DMA Counter Hardware Status structure |
| DMA_Descriptor | DMA_Descriptor Class |
| DMA_DescriptorBase | DMA_Descriptor Class |
| DMA_DirectPutDescriptor | DMA Direct Put Descriptor Class |
| DMA_Fifo_t | Software DMA Fifo structure |
| DMA_FifoHW_t | Hardware DMA Fifo |
| DMA_InjDescriptor_t | DMA Injection Descriptor Structure |
| DMA_InjFifo_t | Injection DMA Fifo Structure |
| DMA_InjFifoGroup_t | DMA Injection Fifo Group Structure |
| DMA_InjFifoRgetFifoFullHandlerEntry_t | Remote Get Fifo Full Handler Table Entry |
| DMA_InjFifoStatus_t | DMA Injection Fifo Status structure |
| DMA_MemoryFifoDescriptor | DMA Memory Fifo Descriptor Class |
| DMA_PacketHeader_t | Torus DMA Hardware Packet Header |
| DMA_PersonalityInfo_t | Static Info from Personality |
| DMA_RecFifo_t | Reception DMA Fifo Structure |
| DMA_RecFifoGroup_t | DMA Reception Fifo Group Structure |
| DMA_RecFifoMap_t | DMA Reception Fifo Map Structure |
| DMA_RecFifoSharedMemory_t | DMA Reception Fifo Shared Memory Structure |
| DMA_RecFifoStatus_t | DMA Reception Fifo Status Structure |
| DMA_RemoteGetDescriptor | DMA Remote Get Descriptor Class |
| DCMF::DMA::DMABaseDesc | |
| DCMF::DMA::DMAManyToMany< TDesc > | |
| DCMF::CDI::DmaMemregionBgpCnk | |
| DCMF::DMA::DMAMulticast< TDesc > | |
| DCMF::DMA::DMAPacketHeader | |
| DCMF::Protocol::MultiSend::DPutMulticast | |
| DCMF::Protocol::MultiSend::DPutMulticastFactory | |
| DCMF::Protocol::MultiSend::DPutRecv | |
| DCMF::DMA::DynamicGetDesc | |
| DCMF::DMA::DynamicMemFifoDesc | |
| DCMF::DMA::DynamicPutDesc | |
| DCMF::Protocol::Send::EagerDataRegistration | Registration class for an eager data message |
| DCMF::Protocol::Send::EagerDMA | Eager non-blocking point-to-point send implementation |
| DCMF::Protocol::Send::EagerDMAFactory | Point-to-point eager send factory |
| DCMF::Protocol::Send::EagerHeadRegistration | Registration class for an eager envelope message |
| DCMF::Protocol::Send::CDI::EagerProtocolFactory< T_Model, T_Device, T_Message > | Eager send protocol factory for CDI packet devices |
| DCMF::Protocol::Send::CDI::EagerRecv< T_Model, T_Device, T_Message > | |
| DCMF::Protocol::Send::EagerRegistration | Registration base class for an eager message |
| DCMF::Protocol::Send::EagerSend | |
| DCMF::Protocol::Send::CDI::envelope_t | |
| CCMI::Executor::Executor | Base Class for all Executors |
| DCMF::Collectives::Broadcast::ExecutorPool | This class implements utilites like a buffer pool and constructs async broadcast queue elements |
| DCMF::Collectives::Allreduce::Factory | Base class for allreduce factory implementations |
| DCMF::Collectives::Allreduce::RectangleRingDput::Factory | Factory class for rectangle ring allreduce implementation |
| DCMF::Collectives::Allreduce::FactoryT< CONNMGR, COMPOSITE > | Factory template class for CompositeT |
| DCMF::Collectives::Allreduce::Tree::FactoryT< CONNMGR, COMPOSITE > | |
| DCMF::FairLockBoxMutex | |
| DCMF::Queueing::Tree::Fifo | |
| fp32_fp32_t | MAXLOC and MINLOC operation element type for float (32-bit) and float (32-bit) data |
| fp32_int32_t | MAXLOC and MINLOC operation element type for float (32-bit) and signed 32-bit data |
| fp64_fp64_t | MAXLOC and MINLOC operation element type for double (64-bit) and double (64-bit) data |
| fp64_int32_t | MAXLOC and MINLOC operation element type for double (64-bit) and signed 32-bit data |
| DCMF::DMA::GenericCounterClient | |
| DCMF::DMA::GenericMemFifoDesc | |
| DCMF::Queueing::Generic::GenericMessage | Interprocess communication message class This message is posted to a Generic device |
| DCMF::Protocol::Send::GenericShortRegistration | |
| DCMF::Protocol::Send::GenericShortSend | |
| DCMF::Collectives::Geometry | |
| DCMF::DMA::DMAMessage::Get | |
| DCMF::DMA::Get< GDesc, TDesc > | |
| DCMF::Protocol::Get::Get | Base class for non-blocking point-to-point get implementations |
| DCMF::Protocol::Get::GetDMA | Default non-blocking point-to-point get implementation |
| DCMF::Protocol::Get::GetDMAFactory | Point-to-point get factory |
| DCMF::Protocol::Get::GetFactory | Base class for point-to-point get factory implementations |
| DCMF::DMA::DMAMessage::GetModel | |
| DCMF::Protocol::Get::GetOverSend | |
| DCMF::Protocol::Get::GetOverSendBase | |
| DCMF::Protocol::Get::GetOverSendFactory | Point-to-point get factory |
| DCMF::Protocol::Get::GetOverSendInfo | |
| DCMF::Protocol::Get::CDI::GetProtocolFactory< T_DmaModel, T_DmaDevice, T_DmaMessage, T_DmaMemregion > | Get protocol factory for CDI dma devices |
| DCMF::DMA::DMAMessage::GetRequest | |
| DCMF::Queueing::GI::giMessage | Global interrupt message class This message is posted to a GI device |
| DCMF::Protocol::MultiSend::GIMulticastFactory | |
| DCMF::Collectives::MultiSend::GIMulticastImpl | |
| CCMI::Schedule::GiSchedule | |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduce | Base class for non-blocking allreduce implementations |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceFactory | Base class for global allreduce factory implementations |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceIntranodeFactory | Global allreduce intranode factory implementation |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceShortLocal | Global Allreduce Short message local nodes |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceShortTreeRecv | Global allreduce short tree-recv nodes |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceShortTreeSend | Global allreduce short tree-send nodes |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceTree | Global allreduce tree |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceTreeFactory | Global allreduce tree factory implementation |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceWQ | Global allreduce workqueue - non-tree-participating ranks |
| DCMF::Device::WorkQueue::GlobalAllreduceWQMessage | |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceWQTreeRecv | Global allreduce workqueue - send tree rank |
| DCMF::Protocol::GlobalAllreduce::GlobalAllreduceWQTreeSend | Global allreduce workqueue - send tree rank |
| DCMF::Protocol::GlobalBarrier::GlobalBarrierFactory | Base class for global barrier factory implementations |
| DCMF::Protocol::GlobalBcast::GlobalBcast | Base class for non-blocking bcast implementations |
| DCMF::Protocol::GlobalBcast::GlobalBcastFactory | Base class for global bcast factory implementations |
| DCMF::Protocol::GlobalBcast::GlobalBcastIntranode | Global bcast intranode |
| DCMF::Protocol::GlobalBcast::GlobalBcastIntranodeFactory | Global bcast intranode factory implementation |
| DCMF::Protocol::GlobalBcast::GlobalBcastTree | Global bcast tree |
| DCMF::Protocol::GlobalBcast::GlobalBcastTreeFactory | Global bcast tree factory implementation |
| DCMF::Protocol::GlobalBcast::GlobalBcastWQ | Global bcast work queue - non-tree ranks |
| DCMF::Device::WorkQueue::GlobalBcastWQMessage | |
| DCMF::Protocol::GlobalBcast::GlobalBcastWQTreeRecv | Global bcast work queue - tree recv rank |
| DCMF::Protocol::GlobalBcast::GlobalBcastWQTreeSend | Global bcast work queue - tree send rank |
| DCMF::Protocol::GlobalAllreduce::GlobalReduceWQ | Global reduce workqueue - non-tree-participating ranks |
| DCMF::Device::WorkQueue::GlobalReduceWQMessage | |
| DCMF::Protocol::GlobalAllreduce::GlobalReduceWQTreeRecv | Global reduce workqueue - recv tree rank |
| DCMF::Protocol::GlobalAllreduce::GlobalReduceWQTreeSend | Global reduce workqueue - send tree rank |
| GlobInt_Status_t | Information returned by status reads. Note that userEnables returns the uptree state in diagnostic read mode |
| DCMF::Protocol::MultiSend::MemFifoMulticast::Head | |
| DCMF::ChannelSys::Info | Used to hold the thread-local global varibles that track the channel state |
| DCMF::DMA::InjFifoGroup | Injection Fifo Group Class |
| int16_int32_t | MAXLOC and MINLOC operation element type for signed 16-bit and signed 32-bit data |
| int32_int32_t | MAXLOC and MINLOC operation element type for signed 32-bit and signed 32-bit data |
| DCMF::Interrupts | Interrupt managment object |
| DCMF::Protocol::GlobalAllreduce::IntranodeAllreduce | Intranode allreduce |
| DCMF::Protocol::GlobalAllreduce::IntranodeReduce | Intranode reduce |
| DCMF::CDI::LinearFifo | |
| DCMF::CDI::LinearFifoPacket | |
| DCMF::CDI::LinearReceptionFifo | |
| DCMF::Protocol::GlobalBarrier::LLSCBarrier | |
| DCMF::Protocol::GlobalBarrier::LLSCFactory | Base class for global barrier factory implementations |
| DCMF::Queueing::Lockbox::LLSCMsg | A local barrier message that takes advantage of the Load Linked and Store Conditional instructions |
| DCMF::Device::WorkQueue::LocalAllreduceWQMessage | |
| DCMF::Device::WorkQueue::LocalBcastWQMessage | |
| DCMF::Device::WorkQueue::LocalReduceWQMessage | |
| LockBox_Barrier_t | Structure contains information required to perform a lockbox barrier. This structure will be filled out by the LockBox_AllocateBarrier() function |
| DCMF::LockBoxAtomic | |
| DCMF::LockBoxBarrier | |
| CCMI::Schedule::LockboxBarrierSchedule | |
| DCMF::LockBoxFactory | |
| DCMF::Queueing::Lockbox::LockboxMessage | Interprocess communication message class This message is posted to a Lockbox device |
| DCMF::LockBoxMutex | |
| DCMF::LockManager | |
| DCMF::Protocol::GlobalBarrier::LockManagerBarrier | |
| DCMF::Protocol::GlobalBarrier::LockManagerBarrierFactory | Base class for global barrier factory implementations |
| DCMF::Queueing::Generic::LockManagerBarrierMsg | A local barrier message that takes advantage of the Load Linked and Store Conditional instructions |
| DCMF::Log::Log | |
| DCMF::Log::LogGroup | Base Class for Logging |
| CCMI::Logging::LogMgr | |
| DCMF::Protocol::MultiSend::LongDataRegistration | Registration Class for Long Head Packet. This class will store persistent data for the payload packets |
| DCMF::Protocol::MultiSend::LongHeadRegistration | Registration Class for Long Head Packet. This class will store persistent data for the head packet |
| DCMF::Protocol::MultiSend::LongMulticast | LongMulticast protocol to multicast messages of arbitrary sizes a list of destinations. It instantiates the DMAMulticast template |
| DCMF::LwarxStwcxNodeAtomic | |
| DCMF::LwarxStwcxNodeBarrier | |
| DCMF::LwarxStwcxNodeMutex | |
| DCMF::LwarxStwcxProcAtomic | |
| DCMF::LwarxStwcxProcMutex | |
| DCMF::Protocol::MultiSend::M2mMessageArg | |
| DCMF::DMA::Collective< TDesc >::M2mPersRecv | |
| mallinfo | |
| malloc_chunk | |
| malloc_params | |
| malloc_segment | |
| malloc_state | |
| malloc_tree_chunk | |
| DCMF::Protocol::MultiSend::Mantissa | |
| DCMF::DMA::Collective< TDesc >::ManyToMany | |
| DCMF::Protocol::MultiSend::ManyToManyFactory | |
| DCMF::Collectives::MultiSend::ManytomanyImpl | Message passing interface suitable for alltoall communiction |
| CCMI::MultiSend::ManytomanyInterface | Message passing interface suitable for alltoall communiction |
| DCMF::Protocol::MultiSend::ManyToManyProtocol | |
| DCMF::Mapping | Coordinate/Address Mapping class This class implements a portability layer for access to the following:
- Torus Hardware Coordinates
- Translation of coordinate to rank
- Translation of rank to coordinate
|
| CCMI::Mapping | |
| CCMI::MatchQueue | |
| CCMI::MatchQueueElem | |
| DCMF::DMA::Collective< TDesc >::McastRecv | |
| DCMF::Protocol::MultiSend::MemFifoHead | Data Type to store protocol information. It is sent in the header packet |
| DCMF::Protocol::MultiSend::MemFifoManyToMany | |
| DCMF::Protocol::MultiSend::MemFifoManyToManyFactory | |
| DCMF::Protocol::MultiSend::MemFifoMulticast | |
| DCMF::Protocol::MultiSend::MemFifoMulticastFactory | |
| DCMF::Protocol::GlobalBarrier::MemoryBarrier | |
| DCMF::Protocol::GlobalBarrier::MemoryBarrierFactory | Base class for global barrier factory implementations |
| DCMF::Queueing::Lockbox::MemoryBarrierMsg | A local barrier message that takes advantage of the Load Linked and Store Conditional instructions |
| DCMF::MemoryManager | Memory Manager class This class implements a portability layer for access to the following:
- Torus device addresses
- Tree device addresses
- GI device addresses
- Scratchpad Window addresses
- Cache Operations
- Virtual to Physical translations
|
| DCMF::Device::WorkQueue::MemoryWorkQueue | Work queue implementation of a contiguous memory buffer |
| DCMF::CDI::DMA::Memregion< T > | Common Device Interface for all dma memory regions |
| DCMF::Memregion | |
| DCMF::DMA::Message | |
| DCMF::Queueing::Message | Base Class for Messages |
| DCMF::CDI::DMA::Message< T > | Common Device Interface for all dma messages |
| DCMF::CDI::Packet::Message< T > | Common Device Interface for all packet messages |
| DCMF::DMA::Collective< TDesc >::MessageArg | |
| DCMF::Protocol::MultiSend::MessageArg | |
| DCMF::Messager | Common messager base class for all platform-specific messager implementations |
| DCMF::Protocol::MultiSend::MFifoManyToManyFactory | |
| DCMF::Protocol::MultiSend::MFifoMcastFactory | The Memory Fifo Multicast Factory |
| DCMF::DMA::DMAManyToMany< TDesc >::Model | Performance optimization to minimize the overhead creation of descriptors. The model stores a pre-cooked descriptor from which a new descriptor can be cloned |
| DCMF::DMA::DMAMessage::Model | Performance optimization to minimize the overhead creation of descriptors |
| DCMF::DMA::DMAMulticast< TDesc >::Model | Performance optimization to minimize the overhead creation of descriptors. The model stores a pre-cooked descriptor from which a new descriptor can be cloned |
| DCMF::CDI::DMA::Model< T_Mod, T_Dev, T_Msg, T_Mem > | |
| DCMF::CDI::Packet::Model< T_Model, T_Device, T_Message > | |
| DCMF::Protocol::MultiSend::MulticastFactory | |
| DCMF::Collectives::MultiSend::MulticastImpl | |
| CCMI::MultiSend::MulticastInterface | Message passing interface suitable for broadcasts, barriers and reductions |
| DCMF::Collectives::Broadcast::MultiColorBroadcastFactoryT< B, afn > | Base factory class for broadcast factory implementations |
| DCMF::Collectives::Broadcast::MultiColorCompositeT< NUMCOLORS, S, sfn, pwcfn > | Base class for synchronous broadcasts |
| DCMF::Mutex | Base Class for Mutex |
| DCMF::Device::WorkQueue::NonVirtWorkQueue | Defines the interfaces and operations for producer/consumer work queues |
| DCMF::CDI::DMA::offset_vec_t | |
| CCMI::Schedule::OneColorRectAllredSched | |
| CCMI::Schedule::OneColorRectangle | |
| CCMI::Schedule::OneColorRectBcastSched | |
| CCMI::Schedule::OneColorRectRedSched | |
| CCMI::Schedule::OneColorTorusRect | |
| P2PTblEntry_t | |
| DCMF::DMA::DMAMulticast< TDesc >::PacketHeader | DMA Packet Header from SPI simplified for message level protocols |
| DCMF::DMA::DMAMessage::PacketHeader | DMA Packet Header from SPI simplified for message level ----------------------------------------------------- |
| DCMF::Device::WorkQueue::PacketWorkQueue | Work queue implementation of a fixed-size shared memory buffer |
| DCMF::Personality | |
| CCMI::ConnectionManager::PhaseBasedConnMgr | |
| CCMI::Executor::PipelinedAllreduce | |
| DCMF::pManagerCallback | |
| DCMF::Protocol::Protocol | Typical base class for most protocol implementations |
| DCMF::Protocol::GlobalAllreduce::Protocol_t | Overlay structure for DCMF_Protocol_t that allows GlobalAllreduce registration to save the actual protocol to be used so we can refer to it later to make optimizations |
| DCMF::Protocol::ProtocolBase | Base class for all protocol implementations |
| DCMF::Protocol::ProtocolFactory | Base class for all protocol factory implementations |
| DCMF::Collectives::ProtocolFactory | Pipeline widths moved to adaptor_pre.h SK (07/08) |
| DCMF::Protocol::Put::Put | Base class for non-blocking point-to-point put implementations |
| DCMF::DMA::DMAMessage::Put | |
| DCMF::Protocol::Put::PutDMA | Default non-blocking point-to-point put implementation |
| DCMF::Protocol::Put::PutDMADirectPutAck | Direct put ack non-blocking point-to-point put implementation |
| DCMF::Protocol::Put::PutDMADirectPutFactory | Point-to-point put factory |
| DCMF::Protocol::Put::PutDMAFactory | Point-to-point put factory |
| DCMF::Protocol::Put::PutDMAMemFifoAck | Memory fifo ack non-blocking point-to-point put implementation |
| DCMF::Protocol::Put::PutDMAMemFifoFactory | Point-to-point put factory |
| DCMF::Protocol::Put::PutFactory | Base class for point-to-point put factory implementations |
| DCMF::DMA::DMAMessage::PutModel | |
| DCMF::Protocol::Put::PutOverSend | |
| DCMF::Protocol::Put::PutOverSendBase | |
| DCMF::Protocol::Put::PutOverSendFactory | Point-to-point put factory |
| DCMF::Protocol::Put::PutOverSendInfo | |
| DCMF::Queueing::Queue | |
| CCMI::Queue | |
| DCMF::Queueing::QueueElem | Base Class for Queue |
| CCMI::QueueElem | Base Class for Queue |
| CCMI::ConnectionManager::RankBasedConnMgr | |
| DCMF::DMA::RecFifoGroup | Reception Fifo Group Class |
| CCMI::Schedule::Rectangle | Structure to store rectangle |
| DCMF::Collectives::Barrier::RectangleLockboxBarrier | Rectangle lockbox barrier |
| DCMF::Collectives::MultiSend::ManytomanyImpl::Recv | |
| DCMF::DMA::DMAMessage::Recv | |
| DCMF::DMA::DMAManyToMany< TDesc >::Recv | Receive class responsible for pulling data out of a memory fifo and copying it to the application buffer. In a manytomany packets will typically come out of order |
| DCMF::DMA::DMAMulticast< TDesc >::Recv | Receive class responsible for pulling data out of a memory fifo and copying it to the application buffer |
| CCMI::Executor::RecvCallbackData_t | Client data for multisend receive done callback |
| DCMF::DMA::DMAManyToMany< TDesc >::Registration | Class to store protocol persisent regisration info, such as the the function ID |
| DCMF::Protocol::MultiSend::Registration | |
| DCMF::DMA::DMAMessage::Registration | Class to store protocol regisration info. Essentially the function ID |
| DCMF::DMA::DMAMulticast< TDesc >::Registration | Class to store protocol regisration info. Essentially static persistent info like function ID, a model to clone descriptors |
| DCMF::Protocol::Send::CDI::RendezvousProtocolFactory< T_PktModel, T_PktDevice, T_PktMessage, T_DmaModel, T_DmaDevice, T_DmaMessage, T_DmaMemregion > | Rendezvous send protocol factory for CDI packet/dma devices |
| DCMF::Protocol::Send::CDI::RendezvousRecv< T_PktModel, T_PktMessage, T_PktDevice, T_DmaModel, T_DmaMessage, T_DmaDevice, T_DmaMemregion > | Rendezvous send recv for CDI packet/dma devices |
| DCMF::Queueing::RequestPool | A pool of request objects |
| DCMF::DMA::ResourceManager | |
| CCMI::Schedule::RingSchedule | |
| DCMF::Protocol::Send::RTS | The Rzv send protcol |
| DCMF::Protocol::Send::CDI::rts_info_t< T_DmaMemregion > | RTS information sent to begin the send operation |
| DCMF::Protocol::Send::RTSPacketInfo | |
| DCMF::Protocol::Send::RzvDMAFactory | |
| DCMF::Protocol::Send::RzvRecv | Class to store receiver messages and state |
| CCMI::Schedule::Schedule | |
| DCMF::DMA::DMAMulticast< TDesc >::Send | Send multicast meassages. Designed for applications such as the Blue Brain Application. The current latency per destination is about 170 cycles which includes the time to inject and create descriptors |
| DCMF::Protocol::Send::Send | Base class for non-blocking point-to-point send implementations |
| DCMF::DMA::DMAMessage::Send | Send a short (single packet) message. On the destination the dispatch function will invoke the application handler on stack |
| DCMF::DMA::DMAManyToMany< TDesc >::Send | Send a many to many (typically alltoall) message |
| CCMI::Executor::SendCallbackData_t | Client data for multisend send done callback |
| CCMI::Executor::Allreduce::SendCallbackData_t | Client data for multisend send done callback |
| DCMF::Protocol::Send::SendFactory | Base class for point-to-point send factory implementations |
| CCMI::Executor::AllreduceBase::SendState | |
| DCMF::Protocol::MultiSend::SharedMem_t | |
| DCMF::Device::WorkQueue::SharedWorkQueue | Work queue implementation of a fixed-size shared memory buffer |
| DCMF::CDI::shmem_fifo_packet_hdr_t | |
| DCMF::CDI::ShmemDevice | |
| DCMF::CDI::ShmemDmaDevice< T_Memregion > | |
| DCMF::CDI::ShmemDmaMessage< T_Memregion > | |
| DCMF::CDI::ShmemDmaMessageBgpCnk | |
| DCMF::CDI::ShmemDmaModelBgpCnk | Common Device Interface for shared memory dma models |
| DCMF::CDI::ShmemFifoPacket | |
| DCMF::CDI::ShmemFifoPacketHeader | |
| DCMF::CDI::ShmemFifos< F, P > | ShmemFifos for the device. explicitly creating the fifos because of atomics usage |
| DCMF::CDI::ShmemPacketMessage | |
| DCMF::CDI::ShmemPacketModel | Common Device Interface for shared memory packet models |
| DCMF::Collectives::Allreduce::Tree::ShmTreeAllreduce | Tree allreduce protocol for SHM mode |
| DCMF::DMA::Short< TDesc > | Send a short (single packet) message. On the destination the dispatch function will invoke the application handler on stack |
| DCMF::Collectives::Allreduce::ShortAsyncCompositeT< SCHEDULE, EXECUTOR > | Allreduce protocol composite |
| DCMF::Collectives::Allreduce::ShortAsyncFactoryT< CONNMGR, COMPOSITE > | Factory template class for ShortCompositeT |
| DCMF::Protocol::Send::ShortDMA | Short non-blocking point-to-point send implementation |
| DCMF::Protocol::MultiSend::ShortMulticast | Short multicast protocol that multicast's data to a list of destinations by instantiating the DMAMulticast template |
| CCMI::Schedule::ShortRectangleSchedule | |
| DCMF::Protocol::Send::ShortRegistration | |
| DCMF::Protocol::MultiSend::ShortRegistration | Registration Class for a short message. This class will store persistent data for a single packet message |
| DCMF::Protocol::Send::ShortSend | |
| CCMI::ConnectionManager::ShortTorusConnMgr | |
| DCMF::Collectives::Broadcast::Tree::ShortTreeBcast | Tree broadcast protocol for short msg in VN mode |
| DCMF::Collectives::Broadcast::Tree::ShortTreeFactory | ShortTreeFactory class for broadcast factory implementations |
| CCMI::ConnectionManager::SimpleConnMgr | |
| SPR_ccr0_t | CCR0: Core Configuration Register 0 |
| SPR_csrr0_t | CSRR0 is an SPR that is used to save machine state on critical interrupts |
| SPR_csrr1_t | CSRR1 CSRR1 is an SPR that is used to save machine state on critical interrupts |
| SPR_dac1_t | DAC1: One of two Data Address Compare registers that specifieds the address upon which DAC (and/or DVC) debug events should occur |
| SPR_dac2_t | DAC2: One of two Data Address Compare registers that specifieds the address upon which DAC (and/or DVC) debug events should occur |
| SPR_dbcr0_t | DBCR0 is an SPR that is used to enable debug modes and events, reset the processor, and control timer operation when debugging |
| SPR_dbcr1_t | DBCR1 is an SPR that is used to configure IAC debug events |
| SPR_dbcr2_t | DBCR2 is an SPR that is used to configure DAC and DVC debug events |
| SPR_dbsr_t | DBSR: The Debug Status Register contains status on debug events as well as information on the type of the most recent reset |
| SPR_dear_t | DEAR contains the address that was referenced by a load, store, or cache management instruction that caused an Alignment, Data TLB Miss, or Data Storage exception |
| SPR_dec_t | DEC: The Decrementer is a 32-bit privileged SPR that decrements at the same rate that the time base increments |
| SPR_iac2_t | IAC2: One of four IAC registers used to specify the addresses upon which IAC debug events should occur |
| SPR_iac3_t | IAC3: One of four IAC registers used to specify the addresses upon which IAC debug events should occur |
| SPR_iac4_t | IAC4: One of four IAC registers used to specify the addresses upon which IAC debug events should occur |
| SPR_iacIAC1t | IAC1: One of four IAC registers used to specify the addresses upon which IAC debug events should occur |
| SPR_icdbdr_t | ICDBDR: Instruction Cache Debug Data Register |
| SPR_icdbtrh_t | ICDBTRH: Instruction Cache Debug Tag Register High |
| SPR_icdbtrl_t | ICDBTRL: Instruction Cache Debug Tag Register Low |
| SPR_ivor0_t | IVOR0 - Critical Interrupt Vector Offset |
| SPR_ivor10_t | IVOR10 - Decrementer Interrupt Vector Offset |
| SPR_ivor11_t | IVOR11 - Fixed Interval Timer Interrupt Vector Offset |
| SPR_ivor12_t | IVOR12 - Watchdog Timer Interrupt Vector Offset |
| SPR_ivor13_t | IVOR13 - Data TLB Error Interrupt Vector Offset |
| SPR_ivor14_t | IVOR14 - Instruction TLB Error Interrupt Vector Offset |
| SPR_ivor15_t | IVOR015 - Debug Interrupt Vector Offset |
| SPR_ivor1_t | IVOR1 - Machine Check Interrupt Vector Offset |
| SPR_ivor2_t | IVOR2 - Data Storage Interrupt Vector Offset |
| SPR_ivor3_t | IVOR3 - Instruction Storage Interrupt Vector Offset |
| SPR_ivor4_t | IVOR4 - External Input Interrrupt Vector Offset |
| SPR_ivor5_t | IVOR5 - Alignment Interrupt Vector Offset |
| SPR_ivor6_t | IVOR6 - Program Interrupt Vector Offset |
| SPR_ivor7_t | IVOR7 - Floating Point Unavailable Interrupt Vector Offset |
| SPR_ivor8_t | IVOR8 - System Call Interrupt Vector Offset |
| SPR_ivor9_t | IVOR9 - Auxiliary Processor Unavailable Interrupt Vector Offset |
| SPR_ivpr_t | IVPR - provides the high-order 16 bits of the effective address of the interrupt vectors, for all interrupt types |
| SPR_mcsr_t | MCSR: The Machine Check Status Register contains status to allow the Machine Check interrupt handler software to determine the cause of a machine check exception |
| SPR_mcsrr0_t | MCSRR0 is an SPR that is used to save machine state on Machine Check interrupts |
| SPR_mcsrr1_t | MCSRR1 is an SPR that is used to save machine state on Machine Check interrupts |
| SPR_mmucr_t | MMUCR: Memory Management Unit Control Register |
| SPR_msr_t | MSR: Controls important chip functions such as enabling/disabling of interrupts |
| SPR_sprg0_sprg7_t | SPRG0-SPRG7: Special Purpose Registers 0 through 7. Provided for general purpose system-dependent software use |
| SPR_srr0_t | SRR0 is an SPR that is used to save machine state on non-critical interrupts |
| SPR_srr1_t | SRR1 is an SPR that is used to save machine state on non-critical interrupts |
| SPR_tcr_t | TCR: The Timer Control Register is a privileged SPR that controls DEC, FIT, and Watchdog Timer operation |
| SPR_tsr_t | TSR: The Timer Status Register is a privileged SPR that records the status of DEC, FIT, and Watchdog Timer events |
| DCMF::Tree::StatusRegister | |
| DCMF::Tree::StatusShadow_tag | |
| DCMF::SysDep | SysDep System Dependency Class This object is a portability layer that abstracts the important features of the low level hardware. This object currently provides the following features:
- Access to the personality object
- Access to the mapping object
- Access to the memory manager object
- Access to the lock manager object
|
| T_BGP_Atomic | |
| T_BGP_Pers_L1Cfg | |
| T_BGP_Pers_L3Cfg | |
| T_BGP_Pers_L3Select | |
| T_BGP_Personality_t | |
| T_BGP_QuadWord | |
| T_BGP_RAS_Detail | |
| T_BGP_RAS_Event | |
| T_BGP_SprgDataI | |
| T_BGP_SprgDST2 | |
| T_BGP_SprgShMem | |
| T_BGP_SprgTextI | |
| T_BGP_TLB_Data | |
| T_BGP_TLB_Entry | |
| T_BGP_TLB_Word0 | |
| T_BGP_TLB_Word1 | |
| T_BGP_TLB_Word2 | |
| T_BGP_TreeCollectiveHdr | |
| T_BGP_TreeFifoStatus | |
| T_BGP_TreeHwHdr | |
| T_BGP_TreePayload | |
| T_BGP_TreePtpHdr | |
| T_BGP_UniversalComponentIdentifier | |
| T_BGP_VMM_RasData | |
| T_BGP_VMM_State | |
| T_BGP_VMM_TLB_Swap | |
| DCMF::Queueing::Tree::Device::task_dispatch | Actual number of threads attached to tasks |
| DCMF::Queueing::Tree::Device::thread_dispatch | Array of helper threads |
| DCMF::Log::Timer | Timer Class Implements a timer primative for statistics gathering |
| DCMF::Topology | |
| CCMI::ConnectionManager::TorusConnMgr | |
| CCMI::TorusMapping | |
| DCMF::Queueing::Tree::Tree2PRecvMessage | Tree Recv Message base class for 2-pass double-sum |
| DCMF::Protocol::MultiSend::TreeAllreduceMessage | Tree Allreduce Send Message base class |
| DCMF::Protocol::MultiSend::TreeAllreducePreMessage | Tree Allreduce Send Message with pre-processing |
| DCMF::Protocol::MultiSend::TreeAllreduceRecvMessage | Tree Allreduce Recv Message |
| DCMF::Protocol::MultiSend::TreeAllreduceRecvPostMessage | Tree Allreduce Recv Message with post-processing |
| DCMF::Protocol::MultiSend::TreeAllreduceSetup | Object to hold info about a Tree Allreduce |
| DCMF::Protocol::MultiSend::TreeAllreduceShortLocalMessage | Tree Allreduce Local Message for Short allreduces |
| DCMF::Protocol::MultiSend::TreeAllreduceShortRecvMessage | Tree Allreduce Recv Message for Short allreduces |
| DCMF::Protocol::MultiSend::TreeAllreduceShortRecvPostMessage | Tree Allreduce Recv Message for Short allreduces with post-processing |
| DCMF::Protocol::MultiSend::TreeAllreduceShortSendMessage | Tree Allreduce Send Message for Short allreduces |
| DCMF::Protocol::MultiSend::TreeAllreduceShortSendPreMessage | Tree Allreduce Send Message for Short allreduces with pre-processing |
| DCMF::Protocol::MultiSend::TreeAllreduceSum2PMessage | Tree Allreduce DOUBLE-SUM 2-Pass Send |
| DCMF::Protocol::MultiSend::TreeAllreduceSum2PMessageV1 | |
| DCMF::Protocol::MultiSend::TreeAllreduceSum2PRecvMessage | Tree Allreduce DOUBLE-SUM 2-Pass Recv |
| DCMF::Protocol::MultiSend::TreeAllreduceSum2PRecvMessageV1 | |
| DCMF::Protocol::MultiSend::TreeAllreduceSum2PWQMessage | Tree Allreduce DOUBLE-SUM 2-Pass Send using WQ for feedback |
| DCMF::Protocol::MultiSend::TreeAllreduceSum2PWQRecvMessage | Tree Allreduce DOUBLE-SUM 2-Pass Recv using WQ for feedback |
| DCMF::Protocol::MultiSend::TreeBarrierMessage | |
| DCMF::Protocol::MultiSend::TreeBarrierRecvMessage | Non Blocking Tree Barrier Receive Message |
| DCMF::Collectives::Broadcast::Tree::TreeBcastComposite | Tree broadcast protocol. We Enhance the MultiColorComposite with a new SyncBcastPost method |
| DCMF::Collectives::Broadcast::Tree::TreeBcastFactory | TreeBcastFactory class for broadcast factory implementations |
| DCMF::Protocol::MultiSend::TreeBcastMessage | Tree Broadcast Send |
| DCMF::Protocol::MultiSend::TreeBcastRecvMessage | Tree Broadcast Recv |
| CCMI::Schedule::TreeBwSchedule | |
| DCMF::Queueing::Tree::TreeMessage | Common, base class of all Tree Device Messages |
| DCMF::Protocol::MultiSend::TreeMulticastFactory | |
| DCMF::Queueing::Tree::TreeRecvMessage | Tree Recv Message base class |
| CCMI::Schedule::TreeSchedule | |
| DCMF::Queueing::Tree::TreeSendMessage | Tree Send Message base class |
| DCMF::Queueing::Tree::TreeWQ2PRecvMessage | Tree Recv Message base class for 2-pass double-sum using WQ as source |
| DCMF::Protocol::MultiSend::TreeWQAllreduceMessage | Tree Allreduce Send Message that consumes from a WQ |
| DCMF::Protocol::MultiSend::TreeWQAllreducePreMessage | Tree Allreduce Send Message that consumes from a WQ and does pre-processing |
| DCMF::Protocol::MultiSend::TreeWQAllreduceRecvMessage | Tree Allreduce Recv Message that produces into a WQ |
| DCMF::Protocol::MultiSend::TreeWQAllreduceRecvPostMessage | Tree Allreduce Recv Message that produces into a WQ and does post-processing |
| DCMF::Protocol::MultiSend::TreeWQBcastMessage | Tree Broadcast Send with WQ as input |
| DCMF::Protocol::MultiSend::TreeWQBcastRecvMessage | Tree Broadcast Recv with WQ as output |
| DCMF::Queueing::Tree::TreeWQRecvMessage | |
| DCMF::Queueing::Tree::TreeWQSendMessage | |
| uint16_int32_t | MAXLOC and MINLOC operation element type for unsigned 16-bit and signed 32-bit data |
| uint32_int32_t | MAXLOC and MINLOC operation element type for unsigned 32-bit and signed 32-bit data |
| uint32_uint32_t | MAXLOC and MINLOC operation element type for unsigned 32-bit and unsigned 32-bit data |
| uint64_int32_t | MAXLOC and MINLOC operation element type for unsigned 64-bit and signed 32-bit data |
| uint64_uint64_t | MAXLOC and MINLOC operation element type for unsigned 64-bit and unsigned 64-bit data |
| DCMF::Collectives::Allreduce::Tree::VnDualShortTreeAllreduce | Tree allreduce protocol for short msg in VN mode |
| DCMF::Device::WorkQueue::WorkQueue | Defines the interfaces and operations for producer/consumer work queues |
| DCMF::Device::WorkQueue::SharedWorkQueue::workqueue_t | Work queue structure in shared memory |
| DCMF::Device::WorkQueue::PacketWorkQueue::workqueue_t | Work queue structure in shared memory |
| DCMF::Device::WorkQueue::WQDevice | |
| DCMF::Device::WorkQueue::WQMessage | |
