Merge pull request #27644 from vespa-engine/toregge/use-provided-memory-allocator-for-large-arrays

Use provided memory allocator for large arrays.

1 files changed, 3 insertions, 0 deletions

diff --git a/searchlib/src/vespa/searchlib/tensor/large_subspaces_buffer_type.h b/searchlib/src/vespa/searchlib/tensor/large_subspaces_buffer_type.h
index 8cce08e9d81..835bc55fd40 100644
--- a/searchlib/src/vespa/searchlib/tensor/large_subspaces_buffer_type.h
+++ b/searchlib/src/vespa/searchlib/tensor/large_subspaces_buffer_type.h
@@ -35,6 +35,9 @@ public:
     void fallback_copy(void *newBuffer, const void *oldBuffer, EntryCount num_entries) override;
     void initialize_reserved_entries(void *buffer, EntryCount reserved_entries) override;
     const vespalib::alloc::MemoryAllocator* get_memory_allocator() const override;
+    vespalib::alloc::Alloc initial_alloc() const noexcept {
+        return _memory_allocator ? vespalib::alloc::Alloc::alloc_with_allocator(_memory_allocator.get()) : vespalib::alloc::Alloc::alloc(0, vespalib::alloc::MemoryAllocator::HUGEPAGE_SIZE);
+    }
 };
 
 }