diff --git a/projects/rocr-runtime/runtime/hsa-runtime/core/util/flag.h b/projects/rocr-runtime/runtime/hsa-runtime/core/util/flag.h
index fcba3536cd..fec4d646ff 100644
--- a/projects/rocr-runtime/runtime/hsa-runtime/core/util/flag.h
+++ b/projects/rocr-runtime/runtime/hsa-runtime/core/util/flag.h
@@ -273,6 +273,9 @@ class Flag {
       else
         fprintf(stderr, "Failed to parse HSA_ASYNCEVENTS_THREAD_PRIORITY");
     }
+
+    var = os::GetEnvVar("HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG");
+    enable_3d_swizzle_ = (var == "1") ? true : false;
   }
 
   void parse_masks(uint32_t maxGpu, uint32_t maxCU) {
@@ -391,6 +394,8 @@ class Flag {
 
   int async_events_thread_priority() const { return async_events_thread_priority_; }
 
+  bool enable_3d_swizzle() const { return enable_3d_swizzle_; }
+
  private:
   bool check_flat_scratch_;
   bool enable_vm_fault_message_;
@@ -423,6 +428,7 @@ class Flag {
   bool dev_mem_queue_;
   uint32_t signal_abort_timeout_;
   int  async_events_thread_priority_;
+  bool enable_3d_swizzle_ = false;
 
   SDMA_OVERRIDE enable_sdma_;
   SDMA_OVERRIDE enable_peer_sdma_;
diff --git a/projects/rocr-runtime/runtime/hsa-runtime/image/addrlib/src/gfx12/gfx12addrlib.cpp b/projects/rocr-runtime/runtime/hsa-runtime/image/addrlib/src/gfx12/gfx12addrlib.cpp
index 1b2e5e563d..d80b221bb4 100644
--- a/projects/rocr-runtime/runtime/hsa-runtime/image/addrlib/src/gfx12/gfx12addrlib.cpp
+++ b/projects/rocr-runtime/runtime/hsa-runtime/image/addrlib/src/gfx12/gfx12addrlib.cpp
@@ -1253,7 +1253,8 @@ void Gfx12Lib::SanityCheckSurfSize(
         // image, at least for single samples) but they still have the same number of mip levels as the
         // parent image.  This disconnect produces false assertions below as the image size doesn't apparently
         // support the specified number of mip levels.
-        ((pIn->flags.hiZHiS == 0) || (pIn->numMipLevels == 1)))
+        ((pIn->flags.hiZHiS == 0) || (pIn->numMipLevels == 1))   &&
+        !(pIn->flags.view3dAs2dArray))
     {
         UINT_32  lastMipSize   = 1;
         UINT_32  dataChainSize = 0;
diff --git a/projects/rocr-runtime/runtime/hsa-runtime/image/image_manager_gfx12.cpp b/projects/rocr-runtime/runtime/hsa-runtime/image/image_manager_gfx12.cpp
index a573d69911..8606f279dc 100644
--- a/projects/rocr-runtime/runtime/hsa-runtime/image/image_manager_gfx12.cpp
+++ b/projects/rocr-runtime/runtime/hsa-runtime/image/image_manager_gfx12.cpp
@@ -48,6 +48,7 @@
 #include <algorithm>
 #include <climits>
 
+#include "core/inc/runtime.h"
 #include "inc/hsa_ext_amd.h"
 #include "core/inc/hsa_internal.h"
 #include "core/util/utils.h"
@@ -681,9 +682,39 @@ uint32_t ImageManagerGfx12::GetAddrlibSurfaceInfoNv(
       break;
 
     case HSA_EXT_IMAGE_GEOMETRY_3D:
-      in.resourceType = ADDR_RSRC_TEX_3D;
-      break;
+      {
+	in.resourceType = ADDR_RSRC_TEX_3D;
+	/*
+	 * 3D swizzle modes on GFX12 enforces alignment
+	 * of the number of slices  to the block depth.
+	 * If numSlices = 3 then the 3 slices are
+	 * interleaved for 3D locality among the 8 slices
+	 * that make up each block. This causes the memory
+	 * footprint to jump from an ideal size of ~12 GB
+	 * to ~32 GB.
+	 * 'enable3DSwizzleMode' flag tests for env variable
+	 * HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG to enable or disable
+	 * 3D swizzle:
+	 * true: Keep view3dAs2dArray = 0 for real 3D interleaving.
+	 * false: Use view3dAs2dArray = 1 to avoid the alignment
+	 *       expansion.
+	 * 2D swizzle modes can lower size overhead but may yield
+	 * suboptimal cache behavior for fully 3D volumetric
+	 * operations.
+	 */
+	bool enable3DSwizzleMode = core::Runtime::runtime_singleton_->flag().enable_3d_swizzle();
+	if (enable3DSwizzleMode)
+	{
+		in.flags.view3dAs2dArray = 0;
+	}
+	else
+	{
+		in.flags.view3dAs2dArray = 1;
+	}
+	break;
+      }
   }
+
   in.flags.texture = 1;
 
   if (tileMode == Image::TileMode::LINEAR)
@@ -779,6 +810,7 @@ uint32_t ImageManagerGfx12::GetAddrlibSurfaceInfoNv(
           minSize = surfaceSize;
           bestSwizzle = (Addr3SwizzleMode) i;
         } else if ((surfaceSize * ratioHigh) <= (minSize * ratioLow)) {
+          minSize = surfaceSize;
           bestSwizzle = (Addr3SwizzleMode) i;
         }
       }