6.18. Stream Memory Operations
This section describes the stream memory operations of the low-level CUDA driver application programming interface.
There are two versions of these APIs, a legacy version and a newer V2 version.
V1:
The V1 API is disabled by default. Users are required to explicitly enable it, e.g. on Linux by passing the kernel module parameter shown below: modprobe nvidia NVreg_EnableStreamMemOPs=1 There is currently no way to enable these operations on other operating systems.
Users can programmatically query whether the device supports these operations with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS.
Support for the CU_STREAM_WAIT_VALUE_NOR flag can be queried with CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_WAIT_VALUE_NOR.
Support for the cuStreamWriteValue64() and cuStreamWaitValue64() functions, as well as for the CU_STREAM_MEM_OP_WAIT_VALUE_64 and CU_STREAM_MEM_OP_WRITE_VALUE_64 flags, can be queried with CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS.
Support for both CU_STREAM_WAIT_VALUE_FLUSH and CU_STREAM_MEM_OP_FLUSH_REMOTE_WRITES requires dedicated platform hardware features and can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_FLUSH_REMOTE_WRITES.
V2:
The V2 APIs are available by default on all platforms.
Support for the CU_STREAM_WAIT_VALUE_NOR flag can be queried with CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_WAIT_VALUE_NOR_V2.
Support for the cuStreamWriteValue64() and cuStreamWaitValue64() functions, as well as for the CU_STREAM_MEM_OP_WAIT_VALUE_64 and CU_STREAM_MEM_OP_WRITE_VALUE_64 flags, can be queried with CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS_V2.
Support for both CU_STREAM_WAIT_VALUE_FLUSH and CU_STREAM_MEM_OP_FLUSH_REMOTE_WRITES requires dedicated platform hardware features and can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_FLUSH_REMOTE_WRITES.
V1 & V2:
Note that all memory pointers passed as parameters to these operations are device pointers. Where necessary a device pointer should be obtained, for example with cuMemHostGetDevicePointer().
None of the operations accepts pointers to managed memory buffers (cuMemAllocManaged).
Warning: Improper use of these APIs may deadlock the application. Synchronization ordering established through these APIs is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by these APIs should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order.
Functions
- CUresult cuStreamBatchMemOp ( CUstream stream, unsigned int count, CUstreamBatchMemOpParams* paramArray, unsigned int flags )
- Batch operations to synchronize the stream via memory operations.
- CUresult cuStreamBatchMemOp_v2 ( CUstream stream, unsigned int count, CUstreamBatchMemOpParams* paramArray, unsigned int flags )
- Batch operations to synchronize the stream via memory operations.
- CUresult cuStreamWaitValue32 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
- Wait on a memory location.
- CUresult cuStreamWaitValue32_v2 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
- Wait on a memory location.
- CUresult cuStreamWaitValue64 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
- Wait on a memory location.
- CUresult cuStreamWaitValue64_v2 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
- Wait on a memory location.
- CUresult cuStreamWriteValue32 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
- Write a value to memory.
- CUresult cuStreamWriteValue32_v2 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
- Write a value to memory.
- CUresult cuStreamWriteValue64 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
- Write a value to memory.
- CUresult cuStreamWriteValue64_v2 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
- Write a value to memory.
Functions
- CUresult cuStreamBatchMemOp ( CUstream stream, unsigned int count, CUstreamBatchMemOpParams* paramArray, unsigned int flags )
-
Batch operations to synchronize the stream via memory operations.
Parameters
- stream
- The stream to enqueue the operations in.
- count
- The number of operations in the array. Must be less than 256.
- paramArray
- The types and parameters of the individual operations.
- flags
- Reserved for future expansion; must be 0.
Description
This is a batch version of cuStreamWaitValue32() and cuStreamWriteValue32(). Batching operations may avoid some performance overhead in both the API call and the device execution versus adding them to the stream in separate API calls. The operations are enqueued in the order they appear in the array.
See CUstreamBatchMemOpType for the full set of supported operations, and cuStreamWaitValue32(), cuStreamWaitValue64(), cuStreamWriteValue32(), and cuStreamWriteValue64() for details of specific operations.
Basic support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS. See related APIs for details on querying support for specific operations.
Note:Warning: Improper use of this API may deadlock the application. Synchronization ordering established through this API is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by this API should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order. For more information, see the Stream Memory Operations section in the programming guide(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWaitValue32, cuStreamWaitValue64, cuStreamWriteValue32, cuStreamWriteValue64, cuMemHostRegister
- CUresult cuStreamBatchMemOp_v2 ( CUstream stream, unsigned int count, CUstreamBatchMemOpParams* paramArray, unsigned int flags )
-
Batch operations to synchronize the stream via memory operations.
Parameters
- stream
- The stream to enqueue the operations in.
- count
- The number of operations in the array. Must be less than 256.
- paramArray
- The types and parameters of the individual operations.
- flags
- Reserved for future expansion; must be 0.
Description
This is a batch version of cuStreamWaitValue32_v2() and cuStreamWriteValue32_v2(). Batching operations may avoid some performance overhead in both the API call and the device execution versus adding them to the stream in separate API calls. The operations are enqueued in the order they appear in the array.
See CUstreamBatchMemOpType for the full set of supported operations, and cuStreamWaitValue32_v2(), cuStreamWaitValue64_v2(), cuStreamWriteValue32_v2(), and cuStreamWriteValue64_v2() for details of specific operations.
See related APIs for details on querying support for specific operations.
Note:Warning: Improper use of this API may deadlock the application. Synchronization ordering established through this API is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by this API should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order. For more information, see the Stream Memory Operations section in the programming guide(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWaitValue32_v2, cuStreamWaitValue64_v2, cuStreamWriteValue32_v2, cuStreamWriteValue64_v2, cuMemHostRegister
- CUresult cuStreamWaitValue32 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
-
Wait on a memory location.
Parameters
- stream
- The stream to synchronize on the memory location.
- addr
- The memory location to wait on.
- value
- The value to compare with the memory location.
- flags
- See CUstreamWaitValue_flags.
Description
Enqueues a synchronization of the stream on the given memory location. Work ordered after the operation will block until the given condition on the memory is satisfied. By default, the condition is to wait for (int32_t)(*addr - value) >= 0, a cyclic greater-or-equal. Other condition types can be specified via flags.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer(). This function cannot be used with managed memory (cuMemAllocManaged).
Support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS.
Support for CU_STREAM_WAIT_VALUE_NOR can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_WAIT_VALUE_NOR.
Note:Warning: Improper use of this API may deadlock the application. Synchronization ordering established through this API is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by this API should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order. For more information, see the Stream Memory Operations section in the programming guide(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWaitValue64, cuStreamWriteValue32, cuStreamWriteValue64, cuStreamBatchMemOp, cuMemHostRegister, cuStreamWaitEvent
- CUresult cuStreamWaitValue32_v2 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
-
Wait on a memory location.
Parameters
- stream
- The stream to synchronize on the memory location.
- addr
- The memory location to wait on.
- value
- The value to compare with the memory location.
- flags
- See CUstreamWaitValue_flags.
Description
Enqueues a synchronization of the stream on the given memory location. Work ordered after the operation will block until the given condition on the memory is satisfied. By default, the condition is to wait for (int32_t)(*addr - value) >= 0, a cyclic greater-or-equal. Other condition types can be specified via flags.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer(). This function cannot be used with managed memory (cuMemAllocManaged).
Support for CU_STREAM_WAIT_VALUE_NOR can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_WAIT_VALUE_NOR_V2.
Note:Warning: Improper use of this API may deadlock the application. Synchronization ordering established through this API is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by this API should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order. For more information, see the Stream Memory Operations section in the programming guide(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWaitValue64_v2, cuStreamWriteValue32_v2, cuStreamWriteValue64_v2, cuStreamBatchMemOp_v2, cuMemHostRegister, cuStreamWaitEvent
- CUresult cuStreamWaitValue64 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
-
Wait on a memory location.
Parameters
- stream
- The stream to synchronize on the memory location.
- addr
- The memory location to wait on.
- value
- The value to compare with the memory location.
- flags
- See CUstreamWaitValue_flags.
Description
Enqueues a synchronization of the stream on the given memory location. Work ordered after the operation will block until the given condition on the memory is satisfied. By default, the condition is to wait for (int64_t)(*addr - value) >= 0, a cyclic greater-or-equal. Other condition types can be specified via flags.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer().
Support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS.
Note:Warning: Improper use of this API may deadlock the application. Synchronization ordering established through this API is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by this API should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order. For more information, see the Stream Memory Operations section in the programming guide(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWaitValue32, cuStreamWriteValue32, cuStreamWriteValue64, cuStreamBatchMemOp, cuMemHostRegister, cuStreamWaitEvent
- CUresult cuStreamWaitValue64_v2 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
-
Wait on a memory location.
Parameters
- stream
- The stream to synchronize on the memory location.
- addr
- The memory location to wait on.
- value
- The value to compare with the memory location.
- flags
- See CUstreamWaitValue_flags.
Description
Enqueues a synchronization of the stream on the given memory location. Work ordered after the operation will block until the given condition on the memory is satisfied. By default, the condition is to wait for (int64_t)(*addr - value) >= 0, a cyclic greater-or-equal. Other condition types can be specified via flags.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer().
Support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS_V2.
Note:Warning: Improper use of this API may deadlock the application. Synchronization ordering established through this API is not visible to CUDA. CUDA tasks that are (even indirectly) ordered by this API should also have that order expressed with CUDA-visible dependencies such as events. This ensures that the scheduler does not serialize them in an improper order. For more information, see the Stream Memory Operations section in the programming guide(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWaitValue32_v2, cuStreamWriteValue32_v2, cuStreamWriteValue64_v2, cuStreamBatchMemOp_v2, cuMemHostRegister, cuStreamWaitEvent
- CUresult cuStreamWriteValue32 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
-
Write a value to memory.
Parameters
- stream
- The stream to do the write in.
- addr
- The device address to write to.
- value
- The value to write.
- flags
- See CUstreamWriteValue_flags.
Description
Write a value to memory. Unless the CU_STREAM_WRITE_VALUE_NO_MEMORY_BARRIER flag is passed, the write is preceded by a system-wide memory fence, equivalent to a __threadfence_system() but scoped to the stream rather than a CUDA thread.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer(). This function cannot be used with managed memory (cuMemAllocManaged).
Support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS.
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWriteValue64, cuStreamWaitValue32, cuStreamWaitValue64, cuStreamBatchMemOp, cuMemHostRegister, cuEventRecord
- CUresult cuStreamWriteValue32_v2 ( CUstream stream, CUdeviceptr addr, cuuint32_t value, unsigned int flags )
-
Write a value to memory.
Parameters
- stream
- The stream to do the write in.
- addr
- The device address to write to.
- value
- The value to write.
- flags
- See CUstreamWriteValue_flags.
Description
Write a value to memory.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer(). This function cannot be used with managed memory (cuMemAllocManaged).
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWriteValue64_v2, cuStreamWaitValue32_v2, cuStreamWaitValue64_v2, cuStreamBatchMemOp_v2, cuMemHostRegister, cuEventRecord
- CUresult cuStreamWriteValue64 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
-
Write a value to memory.
Parameters
- stream
- The stream to do the write in.
- addr
- The device address to write to.
- value
- The value to write.
- flags
- See CUstreamWriteValue_flags.
Description
Write a value to memory. Unless the CU_STREAM_WRITE_VALUE_NO_MEMORY_BARRIER flag is passed, the write is preceded by a system-wide memory fence, equivalent to a __threadfence_system() but scoped to the stream rather than a CUDA thread.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer().
Support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS.
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWriteValue32, cuStreamWaitValue32, cuStreamWaitValue64, cuStreamBatchMemOp, cuMemHostRegister, cuEventRecord
- CUresult cuStreamWriteValue64_v2 ( CUstream stream, CUdeviceptr addr, cuuint64_t value, unsigned int flags )
-
Write a value to memory.
Parameters
- stream
- The stream to do the write in.
- addr
- The device address to write to.
- value
- The value to write.
- flags
- See CUstreamWriteValue_flags.
Description
Write a value to memory.
If the memory was registered via cuMemHostRegister(), the device pointer should be obtained with cuMemHostGetDevicePointer().
Support for this can be queried with cuDeviceGetAttribute() and CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS_V2.
Note:Note that this function may also return error codes from previous, asynchronous launches.
See also:
cuStreamWriteValue32_v2, cuStreamWaitValue32_v2, cuStreamWaitValue64_v2, cuStreamBatchMemOp_v2, cuMemHostRegister, cuEventRecord