diff --git a/src/a2a3/platform/include/aicpu/platform_regs.h b/src/a2a3/platform/include/aicpu/platform_regs.h
index 392f0ecc3..4a4df9f4a 100644
--- a/src/a2a3/platform/include/aicpu/platform_regs.h
+++ b/src/a2a3/platform/include/aicpu/platform_regs.h
@@ -66,17 +66,6 @@ void set_platform_pmu_reg_addrs(uint64_t pmu_regs);
  */
 uint64_t get_platform_pmu_reg_addrs();
 
-/**
- * Set the ACL device ordinal. Latched once per device by simpler_aicpu_init
- * (from InitArgs.device_id) into this resident-SO global; the executor reads it
- * to make the staged orchestration SO filename unique per device so paired dies
- * sharing the preinstall filesystem never collide.
- */
-void set_orch_device_id(int device_id);
-
-/** Get the ACL device ordinal set for the current run (0 if unset). */
-int get_orch_device_id();
-
 #ifdef __cplusplus
 }
 #endif
diff --git a/src/a2a3/platform/shared/aicpu/platform_regs.cpp b/src/a2a3/platform/shared/aicpu/platform_regs.cpp
index b5fc50d1c..055300a14 100644
--- a/src/a2a3/platform/shared/aicpu/platform_regs.cpp
+++ b/src/a2a3/platform/shared/aicpu/platform_regs.cpp
@@ -37,7 +37,6 @@
 
 static uint64_t g_platform_regs = 0;
 static uint64_t g_platform_pmu_reg_addrs = 0;
-static int g_orch_device_id = 0;
 
 void set_platform_regs(uint64_t regs) { g_platform_regs = regs; }
 
@@ -47,10 +46,6 @@ void set_platform_pmu_reg_addrs(uint64_t pmu_regs) { g_platform_pmu_reg_addrs =
 
 uint64_t get_platform_pmu_reg_addrs() { return g_platform_pmu_reg_addrs; }
 
-void set_orch_device_id(int device_id) { g_orch_device_id = device_id; }
-
-int get_orch_device_id() { return g_orch_device_id; }
-
 volatile uint32_t *get_reg_ptr(uint64_t reg_base_addr, RegId reg) {
     return reinterpret_cast<volatile uint32_t *>(reg_base_addr + reg_offset(reg));
 }
diff --git a/src/a2a3/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp b/src/a2a3/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp
index 19f6a7e7c..983d47830 100644
--- a/src/a2a3/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp
+++ b/src/a2a3/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp
@@ -45,6 +45,7 @@
 #include "common/unified_log.h"
 
 // Register-based communication
+#include "aicpu/aicpu_device_config.h"
 #include "aicpu/platform_aicpu_affinity.h"
 #include "aicpu/platform_regs.h"
 #include "common/platform_config.h"
diff --git a/src/a5/platform/include/aicpu/platform_regs.h b/src/a5/platform/include/aicpu/platform_regs.h
index caad59443..9e8e3a635 100644
--- a/src/a5/platform/include/aicpu/platform_regs.h
+++ b/src/a5/platform/include/aicpu/platform_regs.h
@@ -57,17 +57,6 @@ void set_platform_regs(uint64_t regs);
  */
 uint64_t get_platform_regs();
 
-/**
- * Set the ACL device ordinal. Latched once per device by simpler_aicpu_init
- * (from InitArgs.device_id) into this resident-SO global; the executor reads it
- * to make the staged orchestration SO filename unique per device so paired dies
- * sharing the preinstall filesystem never collide.
- */
-void set_orch_device_id(int device_id);
-
-/** Get the ACL device ordinal set for the current run (0 if unset). */
-int get_orch_device_id();
-
 #ifdef __cplusplus
 }
 #endif
diff --git a/src/a5/platform/shared/aicpu/platform_regs.cpp b/src/a5/platform/shared/aicpu/platform_regs.cpp
index 501942fd4..6a74d2a21 100644
--- a/src/a5/platform/shared/aicpu/platform_regs.cpp
+++ b/src/a5/platform/shared/aicpu/platform_regs.cpp
@@ -26,16 +26,11 @@
 #include "spin_hint.h"
 
 static uint64_t g_platform_regs = 0;
-static int g_orch_device_id = 0;
 
 void set_platform_regs(uint64_t regs) { g_platform_regs = regs; }
 
 uint64_t get_platform_regs() { return g_platform_regs; }
 
-void set_orch_device_id(int device_id) { g_orch_device_id = device_id; }
-
-int get_orch_device_id() { return g_orch_device_id; }
-
 void platform_init_aicore_regs(uint64_t reg_addr) {
     // Initialize task dispatch register to idle state
     write_reg(reg_addr, RegId::DATA_MAIN_BASE, AICPU_IDLE_TASK_ID);
diff --git a/src/a5/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp b/src/a5/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp
index 6b8f310da..b40d7fda3 100644
--- a/src/a5/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp
+++ b/src/a5/runtime/tensormap_and_ringbuffer/aicpu/aicpu_executor.cpp
@@ -46,6 +46,7 @@
 #include "common/unified_log.h"
 
 // Register-based communication
+#include "aicpu/aicpu_device_config.h"
 #include "aicpu/platform_regs.h"
 #include "common/platform_config.h"
 
diff --git a/src/common/platform/include/aicpu/aicpu_device_config.h b/src/common/platform/include/aicpu/aicpu_device_config.h
index 38ec22e30..9667c0446 100644
--- a/src/common/platform/include/aicpu/aicpu_device_config.h
+++ b/src/common/platform/include/aicpu/aicpu_device_config.h
@@ -28,6 +28,17 @@
 extern "C" {
 #endif
 
+/**
+ * Set the ACL device ordinal. Latched once per device by simpler_aicpu_init
+ * (from InitArgs.device_id) into this resident-SO global; the AICPU executor
+ * reads it to make the staged orchestration SO filename unique per device so
+ * paired dies sharing the preinstall filesystem never collide.
+ */
+void set_orch_device_id(int device_id);
+
+/** Get the ACL device ordinal set for the current run (0 if unset). */
+int get_orch_device_id();
+
 /**
  * Set the AICPU scheduler no-progress watchdog timeout (ms). Latched once per
  * device by simpler_aicpu_init (from InitArgs.scheduler_timeout_ms); read by
diff --git a/src/common/platform/shared/aicpu/aicpu_device_config.cpp b/src/common/platform/shared/aicpu/aicpu_device_config.cpp
index 4d6d38e91..eea91e456 100644
--- a/src/common/platform/shared/aicpu/aicpu_device_config.cpp
+++ b/src/common/platform/shared/aicpu/aicpu_device_config.cpp
@@ -11,11 +11,16 @@
 #include "aicpu/aicpu_device_config.h"
 
 namespace {
-// Latched once per device by simpler_aicpu_init; survives every per-task launch
+// Latched once per device by simpler_aicpu_init; survive every per-task launch
 // because the AICPU inner SO stays dlopen'd for the runner's life.
+int g_orch_device_id = 0;
 int g_scheduler_timeout_ms = 0;
 }  // namespace
 
+void set_orch_device_id(int device_id) { g_orch_device_id = device_id; }
+
+int get_orch_device_id() { return g_orch_device_id; }
+
 void set_scheduler_timeout_ms(int timeout_ms) { g_scheduler_timeout_ms = timeout_ms; }
 
 int get_scheduler_timeout_ms() { return g_scheduler_timeout_ms; }