Debugger API :: CUDA Toolkit Documentation

4.1. CUDBGAPI_st Struct Reference

Public Variables

CUDBGResult ( *acknowledgeEvent30 )( CUDBGEvent30* event ): Inform the debugger API that the event has been processed.
CUDBGResult ( *acknowledgeEvents42 )( ): Inform the debugger API that synchronous events have been processed.
CUDBGResult ( *acknowledgeSyncEvents )( ): Inform the debugger API that synchronous events have been processed.
CUDBGResult ( *clearAttachState )( ): Clear attach-specific state prior to detach.
CUDBGResult ( *disassemble )( uint32_t dev, uint64_t addr, uint32_t* instSize, char* buf, uint32_t sz ): Disassemble instruction at instruction address.
CUDBGResult ( *finalize )( ): Finalize the API and free all memory.
CUDBGResult ( *getAdjustedCodeAddress )( uint32_t devId, uint64_t address, uint64_t* adjustedAddress, CUDBGAdjAddrAction adjAction ): The client must call this function before inserting a breakpoint, or when the previous or next code address is needed. Returns the adjusted code address for a given code address for a given device.
CUDBGResult ( *getBlockDim )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim3* blockDim ): Get the number of threads in the given block.
CUDBGResult ( *getDevicePCIBusInfo )( uint32_t devId, uint32_t* pciBusId, uint32_t* pciDevId ): Get PCI bus and device ids associated with device devId.
CUDBGResult ( *getDeviceType )( uint32_t dev, char* buf, uint32_t sz ): Get the string description of the device.
CUDBGResult ( *getElfImage )( uint32_t dev, uint32_t sm, uint32_t wp, bool relocated, void* *elfImage, uint64_t* size ): Get the relocated or non-relocated ELF image and size for the grid on the given device.
CUDBGResult ( *getElfImage32 )( uint32_t dev, uint32_t sm, uint32_t wp, bool relocated, void* *elfImage, uint32_t* size ): Get the relocated or non-relocated ELF image and size for the grid on the given device.
CUDBGResult ( *getElfImageByHandle )( uint32_t devId, uint64_t handle, CUDBGElfImageType type, void* elfImage, uint64_t size ): Get the relocated or non-relocated ELF image for the given handle on the given device.
CUDBGResult ( *getGridAttribute )( uint32_t dev, uint32_t sm, uint32_t wp, CUDBGAttribute attr, uint64_t* value ): Get the value of a grid attribute.
CUDBGResult ( *getGridAttributes )( uint32_t dev, uint32_t sm, uint32_t wp, CUDBGAttributeValuePair* pairs, uint32_t numPairs ): Get several grid attribute values in a single API call.
CUDBGResult ( *getGridDim )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim3* gridDim ): Get the number of blocks in the given grid.
CUDBGResult ( *getGridDim32 )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim2* gridDim ): Get the number of blocks in the given grid.
CUDBGResult ( *getGridInfo )( uint32_t dev, uint64_t gridId64, CUDBGGridInfo* gridInfo ): Get information about the specified grid. If the context of the grid has already been destroyed, the function will return CUDBG_ERROR_INVALID_GRID, although the grid id is correct.
CUDBGResult ( *getGridStatus )( uint32_t dev, uint64_t gridId64, CUDBGGridStatus* status ): Check whether the grid corresponding to the given gridId is still present on the device.
CUDBGResult ( *getGridStatus50 )( uint32_t dev, uint32_t gridId, CUDBGGridStatus* status ): Check whether the grid corresponding to the given gridId is still present on the device.
CUDBGResult ( *getHostAddrFromDeviceAddr )( uint32_t dev, uint64_t device_addr, uint64_t* host_addr ): given a device virtual address, return a corresponding system memory virtual address.
CUDBGResult ( *getManagedMemoryRegionInfo )( uint64_t startAddress, CUDBGMemoryInfo* memoryInfo, uint32_t memoryInfo_size, uint32_t* numEntries ): Returns a sorted list of managed memory regions The sorted list of memory regions starts from a region containing the specified starting address. If the starting address is set to 0, a sorted list of managed memory regions is returned which starts from the managed memory region with the lowest start address.
CUDBGResult ( *getNextAsyncEvent50 )( CUDBGEvent50* event ): Copies the next available event in the asynchronous event queue into 'event' and removes it from the queue. The asynchronous event queue is held separate from the normal event queue, and does not require acknowledgement from the debug client.
CUDBGResult ( *getNextAsyncEvent55 )( CUDBGEvent55* event ): Copies the next available event in the asynchronous event queue into 'event' and removes it from the queue. The asynchronous event queue is held separate from the normal event queue, and does not require acknowledgement from the debug client.
CUDBGResult ( *getNextEvent )( CUDBGEventQueueType type, CUDBGEvent* event ): Copies the next available event into 'event' and removes it from the queue.
CUDBGResult ( *getNextEvent30 )( CUDBGEvent30* event ): Copies the next available event in the event queue into 'event' and removes it from the queue.
CUDBGResult ( *getNextEvent32 )( CUDBGEvent32* event ): Copies the next available event in the event queue into 'event' and removes it from the queue.
CUDBGResult ( *getNextEvent42 )( CUDBGEvent42* event ): Copies the next available event in the event queue into 'event' and removes it from the queue.
CUDBGResult ( *getNextSyncEvent50 )( CUDBGEvent50* event )
CUDBGResult ( *getNextSyncEvent55 )( CUDBGEvent55* event ): Copies the next available event in the synchronous event queue into 'event' and removes it from the queue.
CUDBGResult ( *getNumDevices )( uint32_t* numDev ): Get the number of installed CUDA devices.
CUDBGResult ( *getNumLanes )( uint32_t dev, uint32_t* numLanes ): Get the number of lanes per warp on the device.
CUDBGResult ( *getNumRegisters )( uint32_t dev, uint32_t* numRegs ): Get the number of registers per lane on the device.
CUDBGResult ( *getNumSMs )( uint32_t dev, uint32_t* numSMs ): Get the total number of SMs on the device.
CUDBGResult ( *getNumWarps )( uint32_t dev, uint32_t* numWarps ): Get the number of warps per SM on the device.
CUDBGResult ( *getPhysicalRegister30 )( uint64_t pc, char* reg, uint32_t* buf, uint32_t sz, uint32_t* numPhysRegs, CUDBGRegClass* regClass ): Get the physical register number(s) assigned to a virtual register name 'reg' at a given PC, if 'reg' is live at that PC.
CUDBGResult ( *getPhysicalRegister40 )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t pc, char* reg, uint32_t* buf, uint32_t sz, uint32_t* numPhysRegs, CUDBGRegClass* regClass ): Get the physical register number(s) assigned to a virtual register name 'reg' at a given PC, if 'reg' is live at that PC.
CUDBGResult ( *getSmType )( uint32_t dev, char* buf, uint32_t sz ): Get the SM type of the device.
CUDBGResult ( *getTID )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t* tid ): Get the ID of the Linux thread hosting the context of the grid.
CUDBGResult ( *initialize )( ): Initialize the API.
CUDBGResult ( *initializeAttachStub )( ): Initialize the attach stub.
CUDBGResult ( *isDeviceCodeAddress )( uintptr_t addr, bool* isDeviceAddress ): Determines whether a virtual address resides within device code.
CUDBGResult ( *isDeviceCodeAddress55 )( uintptr_t addr, bool* isDeviceAddress ): Determines whether a virtual address resides within device code. This API is strongly deprecated. Use CUDBGAPI_st::isDeviceCodeAddress instead.
CUDBGResult ( *lookupDeviceCodeSymbol )( char* symName, bool* symFound, uintptr_t* symAddr ): Determines whether a symbol represents a function in device code and returns its virtual address.
CUDBGResult ( *memcheckReadErrorAddress )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t* address, ptxStorageKind* storage ): Get the address that memcheck detected an error on.
CUDBGResult ( *readActiveLanes )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t* activeLanesMask ): Reads the bitmask of active lanes on a valid warp.
CUDBGResult ( *readBlockIdx )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim3* blockIdx ): Reads the CUDA block index running on a valid warp.
CUDBGResult ( *readBlockIdx32 )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim2* blockIdx ): Reads the two-dimensional CUDA block index running on a valid warp.
CUDBGResult ( *readBrokenWarps )( uint32_t dev, uint32_t sm, uint64_t* brokenWarpsMask ): Reads the bitmask of warps that are at a breakpoint on a given SM.
CUDBGResult ( *readCallDepth )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t* depth ): Reads the call depth (number of calls) for a given lane.
CUDBGResult ( *readCallDepth32 )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t* depth ): Reads the call depth (number of calls) for a given warp.
CUDBGResult ( *readCodeMemory )( uint32_t dev, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the code memory segment.
CUDBGResult ( *readConstMemory )( uint32_t dev, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the constant memory segment.
CUDBGResult ( *readDeviceExceptionState )( uint32_t devId, uint64_t* exceptionSMMask ): Get the exception state of the SMs on the device.
CUDBGResult ( *readErrorPC )( uint32_t devId, uint32_t sm, uint32_t wp, uint64_t* errorPC, bool* errorPCValid ): Get the hardware reported error PC if it exists.
CUDBGResult ( *readGenericMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t addr, void* buf, uint32_t sz ): Reads content at an address in the generic address space. This function determines if the given address falls into the local, shared, or global memory window. It then accesses memory taking into account the hardware co-ordinates provided as inputs.
CUDBGResult ( *readGlobalMemory )( uint64_t addr, void* buf, uint32_t sz ): Reads content at an address in the global address space. If the address is valid on more than one device and one of those devices does not support UVA, an error is returned.
CUDBGResult ( *readGlobalMemory31 )( uint32_t dev, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the global memory segment.
CUDBGResult ( *readGlobalMemory55 )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the global memory segment (entire 40-bit VA on Fermi+).
CUDBGResult ( *readGridId )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t* gridId64 ): Reads the 64-bit CUDA grid index running on a valid warp.
CUDBGResult ( *readGridId50 )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t* gridId ): Reads the CUDA grid index running on a valid warp.
CUDBGResult ( *readLaneException )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, CUDBGException_t* exception ): Reads the exception type for a given lane.
CUDBGResult ( *readLaneStatus )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, bool* error ): Reads the status of the given lane. For specific error values, use readLaneException.
CUDBGResult ( *readLocalMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the local memory segment.
CUDBGResult ( *readPC )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t* pc ): Reads the PC on the given active lane.
CUDBGResult ( *readParamMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the param memory segment.
CUDBGResult ( *readPinnedMemory )( uint64_t addr, void* buf, uint32_t sz ): Reads content at pinned address in system memory.
CUDBGResult ( *readRegister )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t regno, uint32_t* val ): Reads content of a hardware register.
CUDBGResult ( *readRegisterRange )( uint32_t devId, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t index, uint32_t registers_size, uint32_t* registers ): Reads content of a hardware range of hardware registers.
CUDBGResult ( *readReturnAddress )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t level, uint64_t* ra ): Reads the physical return address for a call level.
CUDBGResult ( *readReturnAddress32 )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t level, uint64_t* ra ): Reads the physical return address for a call level.
CUDBGResult ( *readSharedMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t addr, void* buf, uint32_t sz ): Reads content at address in the shared memory segment.
CUDBGResult ( *readSyscallCallDepth )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t* depth ): Reads the call depth of syscalls for a given lane.
CUDBGResult ( *readTextureMemory )( uint32_t devId, uint32_t vsm, uint32_t wp, uint32_t id, uint32_t dim, uint32_t* coords, void* buf, uint32_t sz ): Read the content of texture memory with given id and coords on sm_20 and lower.
CUDBGResult ( *readTextureMemoryBindless )( uint32_t devId, uint32_t vsm, uint32_t wp, uint32_t texSymtabIndex, uint32_t dim, uint32_t* coords, void* buf, uint32_t sz ): Read the content of texture memory with given symtab index and coords on sm_30 and higher.
CUDBGResult ( *readThreadIdx )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, CuDim3* threadIdx ): Reads the CUDA thread index running on valid lane.
CUDBGResult ( *readValidLanes )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t* validLanesMask ): Reads the bitmask of valid lanes on a given warp.
CUDBGResult ( *readValidWarps )( uint32_t dev, uint32_t sm, uint64_t* validWarpsMask ): Reads the bitmask of valid warps on a given SM.
CUDBGResult ( *readVirtualPC )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t* pc ): Reads the virtual PC on the given active lane.
CUDBGResult ( *readVirtualReturnAddress )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t level, uint64_t* ra ): Reads the virtual return address for a call level.
CUDBGResult ( *readVirtualReturnAddress32 )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t level, uint64_t* ra ): Reads the virtual return address for a call level.
CUDBGResult ( *readWarpState )( uint32_t devId, uint32_t sm, uint32_t wp, CUDBGWarpState* state ): Get state of a given warp.
CUDBGResult ( *requestCleanupOnDetach )( uint32_t appResumeFlag ): Request for cleanup of driver state when detaching.
CUDBGResult ( *requestCleanupOnDetach55 )( ): Request for cleanup of driver state when detaching.
CUDBGResult ( *resumeDevice )( uint32_t dev ): Resume a suspended CUDA device.
CUDBGResult ( *resumeWarpsUntilPC )( uint32_t devId, uint32_t sm, uint64_t warpMask, uint64_t virtPC ): Inserts a temporary breakpoint at the specified virtual PC, and resumes all warps in the specified bitmask on a given SM. As compared to CUDBGAPI_st::resumeDevice, CUDBGAPI_st::resumeWarpsUntilPC provides finer-grain control by resuming a selected set of warps on the same SM. The main intended usage is to accelerate the single-stepping process when the target PC is known in advance. Instead of single-stepping each warp individually until the target PC is hit, the client can issue this API. When this API is used, errors within CUDA kernels will no longer be reported precisely. In the situation where resuming warps is not possible, this API will return CUDBG_ERROR_WARP_RESUME_NOT_POSSIBLE. The client should then fall back to using CUDBGAPI_st::singleStepWarp or CUDBGAPI_st::resumeDevice.
CUDBGResult ( *setBreakpoint )( uint32_t dev, uint64_t addr ): Sets a breakpoint at the given instruction address for the given device. Before setting a breakpoint, CUDBGAPI_st::getAdjustedCodeAddress should be called to get the adjusted breakpoint address.
CUDBGResult ( *setBreakpoint31 )( uint64_t addr ): Sets a breakpoint at the given instruction address.
CUDBGResult ( *setKernelLaunchNotificationMode )( CUDBGKernelLaunchNotifyMode mode ): Set the launch notification policy.
CUDBGResult ( *setNotifyNewEventCallback )( CUDBGNotifyNewEventCallback callback ): Provides the API with the function to call to notify the debugger of a new application or device event.
CUDBGResult ( *setNotifyNewEventCallback31 )( CUDBGNotifyNewEventCallback31 callback, void* data ): Provides the API with the function to call to notify the debugger of a new application or device event.
CUDBGResult ( *setNotifyNewEventCallback40 )( CUDBGNotifyNewEventCallback40 callback ): Provides the API with the function to call to notify the debugger of a new application or device event.
CUDBGResult ( *singleStepWarp )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t* warpMask ): Single step an individual warp on a suspended CUDA device.
CUDBGResult ( *singleStepWarp40 )( uint32_t dev, uint32_t sm, uint32_t wp ): Single step an individual warp on a suspended CUDA device.
CUDBGResult ( *suspendDevice )( uint32_t dev ): Suspends a running CUDA device.
CUDBGResult ( *unsetBreakpoint )( uint32_t dev, uint64_t addr ): Unsets a breakpoint at the given instruction address for the given device.
CUDBGResult ( *unsetBreakpoint31 )( uint64_t addr ): Unsets a breakpoint at the given instruction address.
CUDBGResult ( *writeGenericMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t addr, const void* buf, uint32_t sz ): Writes content to an address in the generic address space. This function determines if the given address falls into the local, shared, or global memory window. It then accesses memory taking into account the hardware co-ordinates provided as inputs.
CUDBGResult ( *writeGlobalMemory )( uint64_t addr, const void* buf, uint32_t sz ): Writes content to an address in the global address space. If the address is valid on more than one device and one of those devices does not support UVA, an error is returned.
CUDBGResult ( *writeGlobalMemory31 )( uint32_t dev, uint64_t addr, const void* buf, uint32_t sz ): Writes content to address in the global memory segment.
CUDBGResult ( *writeGlobalMemory55 )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t addr, const void* buf, uint32_t sz ): Writes content to address in the global memory segment (entire 40-bit VA on Fermi+).
CUDBGResult ( *writeLocalMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint64_t addr, const void* buf, uint32_t sz ): Writes content to address in the local memory segment.
CUDBGResult ( *writeParamMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t addr, const void* buf, uint32_t sz ): Writes content to address in the param memory segment.
CUDBGResult ( *writePinnedMemory )( uint64_t addr, const void* buf, uint32_t sz ): Writes content to pinned address in system memory.
CUDBGResult ( *writeRegister )( uint32_t dev, uint32_t sm, uint32_t wp, uint32_t ln, uint32_t regno, uint32_t val ): Writes content to a hardware register.
CUDBGResult ( *writeSharedMemory )( uint32_t dev, uint32_t sm, uint32_t wp, uint64_t addr, const void* buf, uint32_t sz ): Writes content to address in the shared memory segment.

Variables

CUDBGResult ( *CUDBGAPI_st::acknowledgeEvent30 )( CUDBGEvent30* event )

Inform the debugger API that the event has been processed. Since CUDA 3.0.

Parameters

event: - pointer to the event that has been processed

Returns

CUDBG_SUCCESS

CUDBGResult ( *CUDBGAPI_st::acknowledgeEvents42 )( )

Inform the debugger API that synchronous events have been processed. Since CUDA 3.1.

Returns

CUDBG_SUCCESS

CUDBGResult ( *CUDBGAPI_st::acknowledgeSyncEvents )( )

Inform the debugger API that synchronous events have been processed. Since CUDA 5.0.

Returns

CUDBG_SUCCESS

CUDBGResult ( *CUDBGAPI_st::clearAttachState )( )

Clear attach-specific state prior to detach. Since CUDA 5.0.

Returns

CUDBG_SUCCESS

CUDBGResult ( *CUDBGAPI_st::disassemble )( uint32_t dev, uint64_t addr, uint32_t* instSize, char* buf, uint32_t sz )

Disassemble instruction at instruction address. Since CUDA 3.0.

Parameters

dev: - device index
addr: - instruction address
instSize: - instruction size (32 or 64 bits)
buf: - disassembled instruction buffer
sz: - disassembled instruction buffer size

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_DEVICE, CUDBG_ERROR_UNKNOWN

CUDBGResult ( *CUDBGAPI_st::finalize )( )

Finalize the API and free all memory. Since CUDA 3.0.

See also:

initialize

Returns

CUDBG_SUCCESS, CUDBG_ERROR_UNINITIALIZED, CUDBG_ERROR_COMMUNICATION_FAILURE, CUDBG_ERROR_UNKNOWN

CUDBGResult ( *CUDBGAPI_st::getAdjustedCodeAddress )( uint32_t devId, uint64_t address, uint64_t* adjustedAddress, CUDBGAdjAddrAction adjAction )

The client must call this function before inserting a breakpoint, or when the previous or next code address is needed. Returns the adjusted code address for a given code address for a given device. Since CUDA 5.5.

See also:

unsetBreakpoint

Parameters

devId: - the device index
address
adjustedAddress: - adjusted address
adjAction: - whether the adjusted next, previous or current address is needed

Returns

CUDBG_SUCCESS, CUDBG_ERROR_UNINITIALIZED, CUDBG_ERROR_INVALID_ADDRESS, CUDBG_ERROR_INVALID_DEVICE

CUDBGResult ( *CUDBGAPI_st::getBlockDim )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim3* blockDim )

Get the number of threads in the given block. Since CUDA 3.0.

See also:

getGridDim

Parameters

dev: - device index
sm: - SM index
wp: - warp index
blockDim: - the returned number of threads in the block

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getDevicePCIBusInfo )( uint32_t devId, uint32_t* pciBusId, uint32_t* pciDevId )

Get PCI bus and device ids associated with device devId.

Parameters

devId: - the cuda device id
pciBusId: - pointer where corresponding PCI BUS ID would be stored
pciDevId: - pointer where corresponding PCI DEVICE ID would be stored

Returns

CUDBG_ERROR_INVALID_ARGS, CUDBG_SUCCESS, CUDBG_ERROR_INVALID_DEVICE

CUDBGResult ( *CUDBGAPI_st::getDeviceType )( uint32_t dev, char* buf, uint32_t sz )

Get the string description of the device. Since CUDA 3.0.

See also:

getSMType

Parameters

dev: - device index
buf: - the destination buffer
sz: - the size of the buffer

Returns

CUDBG_SUCCESS, CUDBG_ERROR_BUFFER_TOO_SMALL, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_DEVICE, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getElfImage )( uint32_t dev, uint32_t sm, uint32_t wp, bool relocated, void* *elfImage, uint64_t* size )

Get the relocated or non-relocated ELF image and size for the grid on the given device. Since CUDA 4.0.

Parameters

dev: - device index
sm: - SM index
wp: - warp index
relocated: - set to true to specify the relocated ELF image, false otherwise
*elfImage: - pointer to the ELF image
size: - size of the ELF image (64 bits)

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getElfImage32 )( uint32_t dev, uint32_t sm, uint32_t wp, bool relocated, void* *elfImage, uint32_t* size )

Get the relocated or non-relocated ELF image and size for the grid on the given device. Since CUDA 3.0.

Parameters

dev: - device index
sm: - SM index
wp: - warp index
relocated: - set to true to specify the relocated ELF image, false otherwise
*elfImage: - pointer to the ELF image
size: - size of the ELF image (32 bits)

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getElfImageByHandle )( uint32_t devId, uint64_t handle, CUDBGElfImageType type, void* elfImage, uint64_t size )

Get the relocated or non-relocated ELF image for the given handle on the given device. The handle is provided in the ELF Image Loaded notification event.

Since CUDA 6.0.

Parameters

devId: - device index
handle: - elf image handle
type: - type of the requested ELF image
elfImage: - pointer to the ELF image
size

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_DEVICE, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getGridAttribute )( uint32_t dev, uint32_t sm, uint32_t wp, CUDBGAttribute attr, uint64_t* value )

Get the value of a grid attribute. Since CUDA 3.1.

Parameters

dev: - device index
sm: - SM index
wp: - warp index
attr: - the attribute
value: - the returned value of the attribute

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_INVALID_ATTRIBUTE, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getGridAttributes )( uint32_t dev, uint32_t sm, uint32_t wp, CUDBGAttributeValuePair* pairs, uint32_t numPairs )

Get several grid attribute values in a single API call. Since CUDA 3.1.

Parameters

dev: - device index
sm: - SM index
wp: - warp index
pairs: - array of attribute/value pairs
numPairs: - the number of attribute/values pairs in the array

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_INVALID_ATTRIBUTE, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getGridDim )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim3* gridDim )

Get the number of blocks in the given grid. Since CUDA 4.0.

See also:

getBlockDim

Parameters

dev: - device index
sm: - SM index
wp: - warp index
gridDim: - the returned number of blocks in the grid

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getGridDim32 )( uint32_t dev, uint32_t sm, uint32_t wp, CuDim2* gridDim )

Get the number of blocks in the given grid. Since CUDA 3.0.

See also:

getBlockDim

Parameters

dev: - device index
sm: - SM index
wp: - warp index
gridDim: - the returned number of blocks in the grid

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_UNINITIALIZED

CUDBGResult ( *CUDBGAPI_st::getGridInfo )( uint32_t dev, uint64_t gridId64, CUDBGGridInfo* gridInfo )

Get information about the specified grid. If the context of the grid has already been destroyed, the function will return CUDBG_ERROR_INVALID_GRID, although the grid id is correct. Since CUDA 5.5.

Parameters

dev
gridId64
gridInfo: - pointer to a client allocated structure in which grid info will be returned.

Returns

CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_GRID, CUDBG_SUCCESS

CUDBGResult ( *CUDBGAPI_st::getGridStatus )( uint32_t dev, uint64_t gridId64, CUDBGGridStatus* status )

Check whether the grid corresponding to the given gridId is still present on the device. Since CUDA 5.5.

Parameters

dev
gridId64: - 64-bit grid ID
status: - enum indicating whether the grid status is INVALID, PENDING, ACTIVE, SLEEPING, TERMINATED or UNDETERMINED

Returns

CUDBG_ERROR_INVALID_DEVICE, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_INTERNAL

CUDBGResult ( *CUDBGAPI_st::getGridStatus50 )( uint32_t dev, uint32_t gridId, CUDBGGridStatus* status )

Check whether the grid corresponding to the given gridId is still present on the device. Since CUDA 5.0.

Parameters

dev
gridId: - grid ID
status: - enum indicating whether the grid status is INVALID, PENDING, ACTIVE, SLEEPING, TERMINATED or UNDETERMINED

Returns

CUDBG_ERROR_INVALID_DEVICE, CUDBG_ERROR_INVALID_GRID, CUDBG_ERROR_INTERNAL

CUDBGResult ( *CUDBGAPI_st::getHostAddrFromDeviceAddr )( uint32_t dev, uint64_t device_addr, uint64_t* host_addr )

given a device virtual address, return a corresponding system memory virtual address. Since CUDA 4.1.

See also:

readGenericMemory

writeGenericMemory

Parameters

dev: - device index
device_addr: - device memory address
host_addr: - returned system memory address

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_DEVICE, CUDBG_ERROR_INVALID_CONTEXT, CUDBG_ERROR_INVALID_MEMORY_SEGMENT

CUDBGResult ( *CUDBGAPI_st::getManagedMemoryRegionInfo )( uint64_t startAddress, CUDBGMemoryInfo* memoryInfo, uint32_t memoryInfo_size, uint32_t* numEntries )

Returns a sorted list of managed memory regions The sorted list of memory regions starts from a region containing the specified starting address. If the starting address is set to 0, a sorted list of managed memory regions is returned which starts from the managed memory region with the lowest start address. Since CUDA 6.0.

Parameters

startAddress: - The address that the first region in the list must contain. If the starting address is set to 0, the list of managed memory regions returned starts from the managed memory region with the lowest start address.
memoryInfo: - Client-allocated array of memory region records of type CUDBGMemoryInfo.
memoryInfo_size: - Number of records of type CUDBGMemoryInfo that memoryInfo can hold.
numEntries: - Pointer to a client-allocated variable holding the number of valid entries retured in memoryInfo. Valid entries are continguous and start from memoryInfo[0].

Returns

CUDBG_SUCCESS, CUDBG_ERROR_INVALID_ARGS, CUDBG_ERROR_INVALID_ADDRESS, CUDBG_ERROR_INTERNAL

CUDBGResult ( *CUDBGAPI_st::getNextAsyncEvent50 )( CUDBGEvent50* event )

Copies the next available event in the asynchronous event queue into 'event' and removes it from the queue. The asynchronous event queue is held separate from the normal event queue, and does not require acknowledgement from the debug client. Since CUDA 5.0.