diff --git a/.github/workflows/build-and-test.yml b/.github/workflows/build-and-test.yml
index 50d913e9e..dea5a1bec 100644
--- a/.github/workflows/build-and-test.yml
+++ b/.github/workflows/build-and-test.yml
@@ -12,7 +12,7 @@ jobs:
       fail-fast: false
       matrix:
         build: [
-          {runs-on: ubuntu-latest, enable_perf: OFF, name: "run", ttrt_flags: "--non-zero"},
+          {runs-on: ubuntu-latest, enable_perf: OFF, name: "run", ttrt_flags: ""},
           {runs-on: ubuntu-latest, enable_perf: ON, name: "perf", ttrt_flags: ""},
         ]
 
@@ -147,10 +147,10 @@ jobs:
       fail-fast: false
       matrix:
         build: [
-          {runs-on: n150, enable_perf: OFF, name: "run"},
-          {runs-on: n150, enable_perf: ON, name: "perf"},
-          {runs-on: n300, enable_perf: OFF, name: "run"},
-          {runs-on: n300, enable_perf: ON, name: "perf"},
+          {runs-on: n150, enable_perf: OFF, name: "run", ttrt_flags: "--non-zero"},
+          {runs-on: n150, enable_perf: ON, name: "perf", ttrt_flags: "--non-zero"},
+          {runs-on: n300, enable_perf: OFF, name: "run", ttrt_flags: "--non-zero"},
+          {runs-on: n300, enable_perf: ON, name: "perf", ttrt_flags: "--non-zero"},
         ]
 
     runs-on:
diff --git a/runtime/lib/ttnn/operations/layout/from_device.cpp b/runtime/lib/ttnn/operations/layout/from_device.cpp
index 499b0372e..613859135 100644
--- a/runtime/lib/ttnn/operations/layout/from_device.cpp
+++ b/runtime/lib/ttnn/operations/layout/from_device.cpp
@@ -16,7 +16,14 @@ void run(const ::tt::target::ttnn::FromDeviceOp *op, ProgramContext &context) {
 
   ::ttnn::Tensor out = ::ttnn::from_device(inputTensor);
 
-  tensorPool.try_emplace(op->out()->global_id(), out);
+  if (tensorPool.contains(op->out()->global_id())) {
+    ::ttnn::Tensor &outputTensor = tensorPool.at(op->out()->global_id());
+    void *src = ::tt::tt_metal::get_raw_host_data_ptr(out);
+    void *dst = ::tt::tt_metal::get_raw_host_data_ptr(outputTensor);
+    std::uint32_t size = out.volume() * out.element_size();
+    std::memcpy(dst, src, size);
+  } else {
+    tensorPool.insert_or_assign(op->out()->global_id(), out);
+  }
 }
-
 } // namespace tt::runtime::ttnn::operations::layout