diff --git a/rocclr/device/rocm/rocmemory.cpp b/rocclr/device/rocm/rocmemory.cpp
index 664bc62b63..0e0b021c99 100644
--- a/rocclr/device/rocm/rocmemory.cpp
+++ b/rocclr/device/rocm/rocmemory.cpp
@@ -352,6 +352,7 @@ bool Memory::pinSystemMemory(void* hostPtr, size_t size) {
 }
 
 void Memory::syncCacheFromHost(VirtualGPU& gpu, device::Memory::SyncFlags syncFlags) {
+  amd::ScopedLock lock(owner()->lockMemoryOps());
   // If the last writer was another GPU, then make a writeback
   if (!isHostMemDirectAccess() && (owner()->getLastWriter() != nullptr) &&
       (&dev() != owner()->getLastWriter())) {
@@ -407,7 +408,6 @@ void Memory::syncCacheFromHost(VirtualGPU& gpu, device::Memory::SyncFlags syncFl
         syncFlagsTmp.skipEntire_ = syncFlags.skipEntire_;
       }
 
-      amd::ScopedLock lock(owner()->lockMemoryOps());
       for (auto& sub : owner()->subBuffers()) {
         //! \note Don't allow subbuffer's allocation in the worker thread.
         //! It may cause a system lock, because possible resource
@@ -462,6 +462,8 @@ void Memory::syncCacheFromHost(VirtualGPU& gpu, device::Memory::SyncFlags syncFl
       }
     }
 
+    gpu.releaseGpuMemoryFence();
+
     // Should never fail
     assert(result && "Memory synchronization failed!");
   }