GenericBufferBase

Fully qualified name: isaacsim::core::includes::GenericBufferBase

template<typename T>
class GenericBufferBase : public isaacsim::core::includes::Buffer<T>

Device-generic (CPU or CUDA device) memory buffer implementation.

Manages a buffer of memory allocated in host RAM (using std::vector) or on a CUDA device. Provides functionality for:

• Memory allocation and deallocation.

• Device selection and switching (with or without data transfer).

• Memory copying between host and device.

• Fill the buffer with a constant value.

• Generate a string representation of the buffer.

Note

Uses RAII principles for automatic resource management.

Warning

Requires proper CUDA environment setup and Unified Virtual Address (UVA) space.

Template Parameters:

T – The data type stored in the buffer.

Public Functions

inline GenericBufferBase(const size_t &size = 0, const int device = -1)

Construct a new device-generic (CPU or CUDA device) buffer.

Parameters:

• size – [in] Initial size of the buffer in elements.

• device – [in] CUDA device ID (higher or equal to 0) to allocate on or -1 for CPU.

inline virtual ~GenericBufferBase()

Destructor that ensures proper cleanup of allocated memory.

inline virtual T *data() const

Gets a pointer to the buffer’s data.

Returns:

Pointer to the buffer’s data.

inline virtual size_t size() const

Gets the current size of the buffer.

Returns:

Number of elements in the buffer.

inline virtual void clear()

Clear the buffer.

Clear the buffer by freeing the memory and resetting the size to 0.

inline virtual bool setDevice(

const int device = -1,

bool keepData = false,

)

Changes the device for this buffer.

If the device changes, existing memory is freed on the old device and reallocated on the new device.

Parameters:

• device – [in] New CUDA device ID (higher or equal to 0) or -1 for CPU.

• keepData – [in] If true, the data from the old device is kept on the new device.

Returns:

True if the device was changed, false otherwise.

inline virtual void resize(size_t size)

Resizes the device buffer (synchronous version of resizeAsync).

Reallocates memory if the new size is different from the current size. Handles deallocation of existing memory if necessary.

Parameters:

size – [in] New size in number of elements.

Returns:

True if the size was changed, false otherwise.

inline virtual void resizeAsync(

size_t size,

cudaStream_t cudaStream = 0,

)

Resizes the device buffer asynchronously (asynchronous version of resize).

Reallocates memory if the new size is different from the current size. Handles deallocation of existing memory if necessary.

Parameters:

• size – [in] New size in number of elements.

• cudaStream – [in] CUDA stream to use for the operation.

Returns:

True if the size was changed, false otherwise.

inline virtual void copyTo(void *dst, size_t size)

Copies the buffer’s data to a destination memory (synchronous version of copyToAsync).

Note

If the specified size is greater than the buffer size, only the first .size() elements are copied. A warning message is logged in this case.

Parameters:

• dst – [out] Destination pointer to copy to.

• size – [in] Number of elements to copy.

inline virtual void copyToAsync(

void *dst,

size_t size,

cudaStream_t cudaStream = 0,

)

Copies the buffer’s data to a destination memory asynchronously (asynchronous version of copyTo).

Note

If the specified size is greater than the buffer size, only the first .size() elements are copied. A warning message is logged in this case.

Parameters:

• dst – [out] Destination pointer to copy to.

• size – [in] Number of elements to copy.

• cudaStream – [in] CUDA stream to use for the operation.

inline virtual void copyFrom(const void *src, size_t size)

Copies the buffer’s data from a source memory (synchronous version of copyFromAsync).

Note

If the specified size is greater than the buffer size, only the first .size() elements are copied. A warning message is logged in this case.

Parameters:

• src – [in] Source pointer to copy from.

• size – [in] Number of elements to copy.

inline virtual void copyFromAsync(

const void *src,

size_t size,

cudaStream_t cudaStream = 0,

)

Copies the buffer’s data from a source memory asynchronously (asynchronous version of copyFrom).

Note

If the specified size is greater than the buffer size, only the first .size() elements are copied. A warning message is logged in this case.

Parameters:

• src – [in] Source pointer to copy from.

• size – [in] Number of elements to copy.

• cudaStream – [in] CUDA stream to use for the operation.

inline virtual bool fill(const T &value)

Fills the buffer with a constant value (synchronous version of fillAsync).

Warning

Only data types of size 1, 2, and 4 bytes are supported. Using other data type sizes will log a warning.

Parameters:

value – [in] Value to fill the buffer with.

Returns:

True if the fill was successful, false otherwise.

inline virtual bool fillAsync(

const T &value,

cudaStream_t cudaStream = 0,

)

Fills the buffer with a constant value asynchronously (asynchronous version of fill).

Warning

Only data types of size 1, 2, and 4 bytes are supported. Using other data type sizes will log a warning.

Parameters:

• value – [in] Value to fill the buffer with.

• cudaStream – [in] CUDA stream to use for the operation.

Returns:

True if the fill was successful, false otherwise.

inline std::string toString() const

Generates a string representation of the buffer.

Returns:

String representation of the buffer.

inline size_t sizeofType() const

Gets the size of a single element in bytes.

Returns:

Size of type T in bytes

inline size_t sizeInBytes() const

Gets the total size of the buffer in bytes.

Returns:

Total size of the buffer in bytes

inline MemoryType type() const

Gets the memory type of the buffer.

Returns:

Memory type (Host or Device)

Protected Attributes

MemoryType m_memoryType

Type of memory where the buffer resides.