diff options
| author | Matthew Bentham <matthew.bentham@arm.com> | 2022-11-23 18:17:48 +0000 |
|---|---|---|
| committer | TeresaARM <teresa.charlinreyes@arm.com> | 2022-12-15 12:24:27 +0000 |
| commit | 6b5f674aad30a3438c295c25b5d115007e80b757 (patch) | |
| tree | 3932603e408330c9f5d09b19a4b224e47e996dec /src/backends/reference/RefTensorHandle.hpp | |
| parent | da6bf9e2eac374cd92147d3c60a8af8bd6bc5a37 (diff) | |
| download | armnn-6b5f674aad30a3438c295c25b5d115007e80b757.tar.gz | |
Change the semantics of RefTensorHandle::Import to 'overlay' existing memory
This makes it possible to call Import on an Allocated() or memory-managed Tensor,
which is needed for the current implementation of OptimizerOptions::m_ExportEnabled
to work (as the last layer before the OutputLayer needs to be able to Import the
user's OutputTensor, but this is done after other memory allocation).
Signed-off-by: Matthew Bentham <matthew.bentham@arm.com>
Change-Id: I1a885c2da7b1f0f3964ae53b8135b5e96a66614f
Diffstat (limited to 'src/backends/reference/RefTensorHandle.hpp')
| -rw-r--r-- | src/backends/reference/RefTensorHandle.hpp | 3 |
1 files changed, 1 insertions, 2 deletions
diff --git a/src/backends/reference/RefTensorHandle.hpp b/src/backends/reference/RefTensorHandle.hpp index d916b39ed9..b4dedd5e77 100644 --- a/src/backends/reference/RefTensorHandle.hpp +++ b/src/backends/reference/RefTensorHandle.hpp @@ -71,8 +71,7 @@ private: std::shared_ptr<RefMemoryManager> m_MemoryManager; RefMemoryManager::Pool* m_Pool; mutable void* m_UnmanagedMemory; - bool m_Imported; - bool m_IsImportEnabled; + void* m_ImportedMemory; }; } |