diff --git a/third_party/fwkacllib/inc/external/runtime/rt_error_codes.h b/third_party/fwkacllib/inc/external/runtime/rt_error_codes.h
index 67146dbe..9f216a56 100755
--- a/third_party/fwkacllib/inc/external/runtime/rt_error_codes.h
+++ b/third_party/fwkacllib/inc/external/runtime/rt_error_codes.h
@@ -38,6 +38,7 @@ static const int32_t ACL_ERROR_RT_STREAM_NO_CB_REG           = 107015; // callba
 static const int32_t ACL_ERROR_RT_INVALID_MEMORY_TYPE        = 107016; // invalid memory type
 static const int32_t ACL_ERROR_RT_INVALID_HANDLE             = 107017; // invalid handle
 static const int32_t ACL_ERROR_RT_INVALID_MALLOC_TYPE        = 107018; // invalid malloc type
+static const int32_t ACL_ERROR_RT_WAIT_TIMEOUT               = 107019; // wait timeout
 
 static const int32_t ACL_ERROR_RT_FEATURE_NOT_SUPPORT        = 207000; // feature not support
 static const int32_t ACL_ERROR_RT_MEMORY_ALLOCATION          = 207001; // memory allocation error
@@ -50,6 +51,7 @@ static const int32_t ACL_ERROR_RT_NO_EVENT_RESOURCE          = 207007; // no eve
 static const int32_t ACL_ERROR_RT_NO_STREAM_RESOURCE         = 207008; // no stream resource
 static const int32_t ACL_ERROR_RT_NO_NOTIFY_RESOURCE         = 207009; // no notify resource
 static const int32_t ACL_ERROR_RT_NO_MODEL_RESOURCE          = 207010; // no model resource
+static const int32_t ACL_ERROR_RT_NO_CDQ_RESOURCE            = 207011; // no cdq resource
 
 static const int32_t ACL_ERROR_RT_INTERNAL_ERROR             = 507000; // runtime internal error
 static const int32_t ACL_ERROR_RT_TS_ERROR                   = 507001; // ts internel error
@@ -85,9 +87,14 @@ static const int32_t ACL_ERROR_RT_DEBUG_UNREGISTER_FAIL      = 507030; // debug
 static const int32_t ACL_ERROR_RT_LABEL_CONTEXT              = 507031; // label not in current context
 static const int32_t ACL_ERROR_RT_PROGRAM_USE_OUT            = 507032; // program register num use out
 static const int32_t ACL_ERROR_RT_DEV_SETUP_ERROR            = 507033; // device setup error
+static const int32_t ACL_ERROR_RT_VECTOR_CORE_TIMEOUT        = 507034; // vector core timeout
+static const int32_t ACL_ERROR_RT_VECTOR_CORE_EXCEPTION      = 507035; // vector core exception
+static const int32_t ACL_ERROR_RT_VECTOR_CORE_TRAP_EXCEPTION = 507036; // vector core trap exception
+static const int32_t ACL_ERROR_RT_CDQ_BATCH_ABNORMAL         = 507037; // cdq alloc batch abnormal
 
 static const int32_t ACL_ERROR_RT_DRV_INTERNAL_ERROR         = 507899; // drv internal error
 static const int32_t ACL_ERROR_RT_AICPU_INTERNAL_ERROR       = 507900; // aicpu internal error
+static const int32_t ACL_ERROR_RT_SOCKET_CLOSE               = 507901; // hdc disconnect
 
 #ifdef __cplusplus
 }
diff --git a/third_party/fwkacllib/inc/runtime/base.h b/third_party/fwkacllib/inc/runtime/base.h
index 40bc91f7..7fc1cdea 100644
--- a/third_party/fwkacllib/inc/runtime/base.h
+++ b/third_party/fwkacllib/inc/runtime/base.h
@@ -156,7 +156,7 @@ RTS_API rtError_t rtProfilerTrace(uint64_t id, bool notify, uint32_t flags, rtSt
 
 /**
  * @ingroup profiling_base
- * @brief ts send keypoint for step info.
+ * @brief ts send keypoint profiler log.
  */
 RTS_API rtError_t rtProfilerTraceEx(uint64_t id, uint64_t modelId, uint16_t tagId, rtStream_t stream);
 
@@ -206,7 +206,7 @@ RTS_API rtError_t rtRegDeviceStateCallback(const char *regName, rtDeviceStateCal
 
 /**
  * @ingroup dvrt_base
- * @brief register callback for fail task 
+ * @brief register callback for fail task
  * @param [in] uniName unique register name, can't be null
  * @param [in] callback fail task callback function
  * @param [out] NA
@@ -345,11 +345,11 @@ RTS_API rtError_t rtLabelCreateEx(rtLabel_t *label, rtStream_t stream);
  * @return RT_ERROR_NONE for ok
  * @return RT_ERROR_INVALID_VALUE for error input
  */
-rtError_t rtLabelCreateExV2(rtLabel_t *label, rtModel_t model, rtStream_t stream);
+RTS_API rtError_t rtLabelCreateExV2(rtLabel_t *label, rtModel_t model, rtStream_t stream);
 
 /**
  * @ingroup dvrt_base
- * @brief get current thread last stream id and task id 
+ * @brief get current thread last stream id and task id
  * @param [out] stream id and task id
  * @param [in] null
  * @return RT_ERROR_NONE for ok
diff --git a/third_party/fwkacllib/inc/runtime/config.h b/third_party/fwkacllib/inc/runtime/config.h
index ee104693..c1327c45 100644
--- a/third_party/fwkacllib/inc/runtime/config.h
+++ b/third_party/fwkacllib/inc/runtime/config.h
@@ -46,6 +46,12 @@ typedef enum tagRtChipType {
     CHIP_END,
 } rtChipType_t;
 
+typedef enum tagRtAicpuScheType {
+    SCHEDULE_SOFTWARE = 0, /* Software Schedule */
+    SCHEDULE_SOFTWARE_OPT,
+    SCHEDULE_HARDWARE, /* HWTS Schedule */
+} rtAicpuScheType;
+
 typedef enum tagRtVersion {
     VER_BEGIN = 0,
     VER_NA = VER_BEGIN,
@@ -65,6 +71,7 @@ typedef enum tagRtPlatformType {
     PLATFORM_LHISI_CS,
     PLATFORM_DC,
     PLATFORM_CLOUD_V2,
+    PLATFORM_LHISI_SD3403,
     PLATFORM_END,
 } rtPlatformType_t;
 
@@ -126,6 +133,11 @@ typedef struct tagRtPlatformConfig {
     uint32_t platformConfig;
 } rtPlatformConfig_t;
 
+typedef enum tagRTTaskTimeoutType {
+    RT_TIMEOUT_TYPE_OP_WAIT = 0,
+    RT_TIMEOUT_TYPE_OP_EXECUTE,
+} rtTaskTimeoutType_t;
+
 /**
  * @ingroup
  * @brief get AI core count
@@ -184,6 +196,37 @@ RTS_API rtError_t rtMemGetL2Info(rtStream_t stream, void **ptr, uint32_t *size);
  */
 RTS_API rtError_t rtGetRuntimeVersion(uint32_t *runtimeVersion);
 
+
+/**
+ * @ingroup
+ * @brief get device feature ability by device id, such as task schedule ability.
+ * @param [in] deviceId
+ * @param [in] moduleType
+ * @param [in] featureType
+ * @param [out] value
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ */
+RTS_API rtError_t rtGetDeviceCapability(int32_t deviceId, int32_t moduleType, int32_t featureType, int32_t *value);
+
+/**
+ * @ingroup
+ * @brief set event wait task timeout time.
+ * @param [in] timeout
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ */
+RTS_API rtError_t rtSetOpWaitTimeOut(uint32_t timeout);
+
+/**
+ * @ingroup
+ * @brief set op execute task timeout time.
+ * @param [in] timeout
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ */
+RTS_API rtError_t rtSetOpExecuteTimeOut(uint32_t timeout);
+
 #if defined(__cplusplus) && !defined(COMPILE_OMG_PACKAGE)
 }
 #endif
diff --git a/third_party/fwkacllib/inc/runtime/dev.h b/third_party/fwkacllib/inc/runtime/dev.h
index e82ec5fa..2cf6712f 100644
--- a/third_party/fwkacllib/inc/runtime/dev.h
+++ b/third_party/fwkacllib/inc/runtime/dev.h
@@ -63,6 +63,11 @@ typedef enum tagRtFeatureType {
     FEATURE_TYPE_RSV
 } rtFeatureType_t;
 
+typedef enum tagRtDeviceFeatureType {
+  FEATURE_TYPE_SCHE,
+  FEATURE_TYPE_END,
+} rtDeviceFeatureType_t;
+
 typedef enum tagMemcpyInfo {
     MEMCPY_INFO_SUPPORT_ZEROCOPY = 0,
     MEMCPY_INFO_RSV
diff --git a/third_party/fwkacllib/inc/runtime/event.h b/third_party/fwkacllib/inc/runtime/event.h
index 41e611ea..57948c47 100644
--- a/third_party/fwkacllib/inc/runtime/event.h
+++ b/third_party/fwkacllib/inc/runtime/event.h
@@ -23,12 +23,23 @@
 extern "C" {
 #endif
 
+typedef enum rtEventWaitStatus {
+    EVENT_STATUS_COMPLETE = 0,
+    EVENT_STATUS_NOT_READY = 1,
+    EVENT_STATUS_MAX = 2,
+} rtEventWaitStatus_t;
+
 /**
  * @ingroup event_flags
  * @brief event op bit flags
  */
-#define RT_EVENT_DEFAULT (0x00)
-#define RT_EVENT_WITH_FLAG (0x01)
+#define RT_EVENT_DEFAULT (0x0E)
+#define RT_EVENT_WITH_FLAG (0x0B)
+
+#define RT_EVENT_DDSYNC_NS    0x01U
+#define RT_EVENT_STREAM_MARK  0x02U
+#define RT_EVENT_DDSYNC       0x04U
+#define RT_EVENT_TIME_LINE    0x08U
 
 /**
  * @ingroup dvrt_event
@@ -104,6 +115,16 @@ RTS_API rtError_t rtEventSynchronize(rtEvent_t event);
  */
 RTS_API rtError_t rtEventQuery(rtEvent_t event);
 
+/**
+ * @ingroup dvrt_event
+ * @brief Queries an event's wait status
+ * @param [in] event   event to query
+ * @param [in out] EVENT_WAIT_STATUS status
+ * @return EVENT_STATUS_COMPLETE for complete
+ * @return EVENT_STATUS_NOT_READY for not complete
+ */
+RTS_API rtError_t rtEventQueryWaitStatus(rtEvent_t event, rtEventWaitStatus_t *status);
+
 /**
  * @ingroup dvrt_event
  * @brief computes the elapsed time between events.
@@ -176,6 +197,18 @@ RTS_API rtError_t rtNotifyRecord(rtNotify_t notify, rtStream_t stream);
  */
 RTS_API rtError_t rtNotifyWait(rtNotify_t notify, rtStream_t stream);
 
+/**
+ * @ingroup dvrt_event
+ * @brief Wait for a notify with time out
+ * @param [in] notify_ notify to be wait
+ * @param [in] stream_  input stream
+ * @param [in] timeOut  input timeOut
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ * @return RT_ERROR_STREAM_CONTEXT for stream is not in current ctx
+ */
+RTS_API rtError_t rtNotifyWaitWithTimeOut(rtNotify_t notify_, rtStream_t stream_, uint32_t timeOut);
+
 /**
  * @ingroup dvrt_event
  * @brief Name a notify
diff --git a/third_party/fwkacllib/inc/runtime/kernel.h b/third_party/fwkacllib/inc/runtime/kernel.h
index 402fadef..9b0221c7 100644
--- a/third_party/fwkacllib/inc/runtime/kernel.h
+++ b/third_party/fwkacllib/inc/runtime/kernel.h
@@ -111,6 +111,16 @@ typedef struct rtKernelInfo {
     uint32_t module_size;
 } *rtKernelInfo_t;
 
+/**
+ * @ingroup rt_kernel
+ * @brief op name
+ */
+typedef struct rtKernelLaunchNames {
+    const char *soName;      // defined for so name
+    const char *kernelName;  // defined for kernel type name
+    const char *opName;      // defined for operator name
+} rtKernelLaunchNames_t;
+
 /**
  * @ingroup rt_KernelConfigDump
  * @brief device dump type
@@ -173,13 +183,7 @@ typedef void (*rtCallback_t)(void *fnData);
  * @ingroup rt_kernel
  * @brief magic number of elf binary for aicube
  */
-#define RT_DEV_BINARY_MAGIC_ELF_AICUBE 0x41415247
-
-/**
- * @ingroup rt_kernel
- * @brief magic number of elf binary for aivector
- */
-#define RT_DEV_BINARY_MAGIC_ELF_AIVECTOR 0x41415248
+#define RT_DEV_BINARY_MAGIC_ELF_AICUBE 0x41494343
 
 /**
  * @ingroup rt_kernel_flags
@@ -192,14 +196,14 @@ typedef void (*rtCallback_t)(void *fnData);
 #define RT_KERNEL_CUSTOM_AICPU (0x08)
 
 // STARS topic scheduler sqe : topic_type
-#define RT_KERNEL_DEVICE_FIRST (0X10)
-#define RT_KERNEL_HOST_ONLY (0X20)
-#define RT_KERNEL_HOST_FIRST (0X30)
+#define RT_KERNEL_DEVICE_FIRST (0x10)
+#define RT_KERNEL_HOST_ONLY (0x20)
+#define RT_KERNEL_HOST_FIRST (0x40)
 
 /**
  * @ingroup rt_kernel
  * @brief kernel mode
- */
+**/
 #define RT_DEFAULT_KERNEL_MODE (0x00)
 #define RT_NORMAL_KERNEL_MODE (0x01)
 #define RT_ALL_KERNEL_MODE (0x02)
@@ -222,7 +226,7 @@ RTS_API rtError_t rtDevBinaryRegister(const rtDevBinary_t *bin, void **handle);
 
 /**
  * @ingroup rt_kernel
- * @brief register device binary
+ * @brief register device binary with all kernel
  * @param [in] bin   device binary description
  * @param [out] handle   device binary handle
  * @return RT_ERROR_NONE for ok
@@ -341,7 +345,7 @@ RTS_API rtError_t rtKernelLaunch(const void *stubFunc, uint32_t blockDim, void *
  * @ingroup rt_kernel
  * @brief launch kernel with handle to device
  * @param [in] handle   program
- * @param [in] devFunc    device function description
+ * @param [in] devFunc   device function description.
  * @param [in] blockDim   block dimentions
  * @param [in] args   argments address for kernel function
  * @param [in] argsSize   argements size
@@ -352,7 +356,7 @@ RTS_API rtError_t rtKernelLaunch(const void *stubFunc, uint32_t blockDim, void *
  * @return RT_ERROR_INVALID_VALUE for error input
  */
 RTS_API rtError_t rtKernelLaunchWithHandle(void *handle, const void *devFunc, uint32_t blockDim, void *args, uint32_t argsSize,
-                                           rtSmDesc_t *smDesc, rtStream_t stream, const void *kernelInfo);
+                                            rtSmDesc_t *smDesc, rtStream_t stream_, const void *kernelInfo);
 
 /**
  * @ingroup rt_kernel
@@ -371,7 +375,7 @@ RTS_API rtError_t rtKernelLaunchWithFlag(const void *stubFunc, uint32_t blockDim
                                          rtSmDesc_t *smDesc, rtStream_t stream, uint32_t flags);
 
 /**
- * @ingroup rt_kernel
+ * @ingroup rt_kernel(abandoned)
  * @brief launch kernel to device
  * @param [in] args       argments address for kernel function
  * @param [in] argsSize   argements size
@@ -383,7 +387,21 @@ RTS_API rtError_t rtKernelLaunchWithFlag(const void *stubFunc, uint32_t blockDim
 RTS_API rtError_t rtKernelLaunchEx(void *args, uint32_t argsSize, uint32_t flags, rtStream_t stream);
 
 /**
- * @ingroup rt_kernel
+ * @ingroup rt_kernel(in use)
+ * @brief launch kernel to device
+ * @param [in] opName     opkernel name
+ * @param [in] args       argments address for kernel function
+ * @param [in] argsSize   argements size
+ * @param [in] flags      launch flags
+ * @param [in] stream     associated stream
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ */
+RTS_API rtError_t rtKernelLaunchFwk(const char *opName, void *args, uint32_t argsSize, uint32_t flags,
+                                    rtStream_t rtStream);
+
+/**
+ * @ingroup rt_kernel(abandoned)
  * @brief launch cpu kernel to device
  * @param [in] soName        so name
  * @param [in] kernelName    kernel name
@@ -399,7 +417,22 @@ RTS_API rtError_t rtCpuKernelLaunch(const void *soName, const void *kernelName,
                                     uint32_t argsSize, rtSmDesc_t *smDesc, rtStream_t stream);
 
 /**
- * @ingroup rt_kernel
+ * @ingroup rt_kernel(in use)
+ * @brief launch cpu kernel to device
+ * @param [in] launchNames   names for kernel launch
+ * @param [in] blockDim      block dimentions
+ * @param [in] args          argments address for kernel function
+ * @param [in] argsSize      argments size
+ * @param [in] smDesc        shared memory description
+ * @param [in] stream        associated stream
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ */
+RTS_API rtError_t rtAicpuKernelLaunch(const rtKernelLaunchNames_t *launchNames,
+    uint32_t blockDim, const void *args, uint32_t argsSize, rtSmDesc_t *smDesc, rtStream_t stream);
+
+/**
+ * @ingroup rt_kernel(abandoned)
  * @brief launch cpu kernel to device  with dump identifier
  * @param [in] soName        so name
  * @param [in] kernelName    kernel name
@@ -416,6 +449,22 @@ RTS_API rtError_t rtCpuKernelLaunchWithFlag(const void *soName, const void *kern
                                             const void *args, uint32_t argsSize, rtSmDesc_t *smDesc, rtStream_t stream,
                                             uint32_t flags);
 
+/**
+ * @ingroup rt_kernel(in use)
+ * @brief launch cpu kernel to device  with dump identifier
+ * @param [in] launchNames   names for kernel launch
+ * @param [in] blockDim      block dimentions
+ * @param [in] args          argments address for kernel function
+ * @param [in] argsSize      argments size
+ * @param [in] smDesc        shared memory description
+ * @param [in] stream        associated stream
+ * @param [in] flag          dump flag or others function flag
+ * @return RT_ERROR_NONE for ok
+ * @return RT_ERROR_INVALID_VALUE for error input
+ */
+RTS_API rtError_t rtAicpuKernelLaunchWithFlag(const rtKernelLaunchNames_t *launchNames, uint32_t blockDim,
+    const void *args, uint32_t argsSize, rtSmDesc_t *smDesc, rtStream_t stream, uint32_t flags);
+
 /**
  * @ingroup rt_kernel
  * @brief L1 fusion dump addr transfered to device
diff --git a/third_party/fwkacllib/inc/runtime/mem.h b/third_party/fwkacllib/inc/runtime/mem.h
index 30af85d9..bace4bc6 100644
--- a/third_party/fwkacllib/inc/runtime/mem.h
+++ b/third_party/fwkacllib/inc/runtime/mem.h
@@ -116,6 +116,9 @@ typedef enum tagRtMemInfoType {
 
 typedef enum tagRtRecudeKind {
     RT_MEMCPY_SDMA_AUTOMATIC_ADD = 10,  // D2D, SDMA inline reduce, include 1P, and P2P
+    RT_MEMCPY_SDMA_AUTOMATIC_MAX = 11,
+    RT_MEMCPY_SDMA_AUTOMATIC_MIN = 12,
+    RT_MEMCPY_SDMA_AUTOMATIC_EQUAL = 13,
     RT_RECUDE_KIND_END
 } rtRecudeKind_t;
 
@@ -123,6 +126,14 @@ typedef enum tagRtDataType {
     RT_DATA_TYPE_FP32 = 0,  // fp32
     RT_DATA_TYPE_FP16 = 1,  // fp16
     RT_DATA_TYPE_INT16 = 2, // int16
+    RT_DATA_TYPE_INT4 = 3,  // int4
+    RT_DATA_TYPE_INT8 = 4,  // int8
+    RT_DATA_TYPE_INT32 = 5, // int32
+    RT_DATA_TYPE_BFP16 = 6, // bfp16
+    RT_DATA_TYPE_BFP32 = 7, // bfp32
+    RT_DATA_TYPE_UINT8 = 8, // uint8
+    RT_DATA_TYPE_UINT16= 9, // uint16
+    RT_DATA_TYPE_UINT32= 10,// uint32
     RT_DATA_TYPE_END
 } rtDataType_t;
 
diff --git a/third_party/fwkacllib/inc/runtime/rt_model.h b/third_party/fwkacllib/inc/runtime/rt_model.h
index 30b8f053..a7618b45 100644
--- a/third_party/fwkacllib/inc/runtime/rt_model.h
+++ b/third_party/fwkacllib/inc/runtime/rt_model.h
@@ -135,12 +135,13 @@ typedef struct tagAllKernelTaskInfo {
     uint16_t argsCount;
     uint16_t argsSize;
     uint16_t reserved;
-    const void *dev_func;
+    void *devfunc;
     void *handle;
     uint8_t *smDesc;
     uint8_t *args;
     uint16_t *argsOffset;
 } rtAllKernelTaskInfo_t;
+
 typedef struct tagKernelTaskInfoEx {
     uint32_t flags;
     uint32_t argsSize;
@@ -198,6 +199,13 @@ typedef struct tagProfilerTraceTaskInfo {
     uint32_t reserved[6];
 } rtProfilerTrace_t;
 
+typedef struct tagProfilerTraceExTaskInfo {
+    uint64_t profilerTraceId;
+    uint64_t modelId;
+    uint16_t tagId;
+    uint8_t reserved[22];
+} rtProfilerTraceEx_t;
+
 typedef struct tagrtMemcpyAsyncTaskInfo {
     void *dst;
     uint64_t destMax;
@@ -265,7 +273,7 @@ typedef struct tagTaskInfo {
     union {
         rtKernelTaskInfoEx_t kernelTaskEx;
         rtKernelTaskInfo_t kernelTask;
-        rtAllKernelTaskInfo_t allkernelTask;
+        rtAllKernelTaskInfo_t allKernelTask;
         rtEventTaskInfo_t eventTask;
         rtStreamSwitchTaskInfo_t streamSwitchTask;
         rtStreamActiveTaskInfo_t streamActiveTask;
@@ -273,6 +281,7 @@ typedef struct tagTaskInfo {
         rtLabelSwitchTaskInfo_t labelSwitchTask;
         rtLabelGotoTaskInfo_t labelGotoTask;
         rtProfilerTrace_t profilertraceTask;
+        rtProfilerTraceEx_t profilertraceExTask;
         rtMemcpyAsyncTaskInfo_t memcpyAsyncTask;
         rtNotifyTaskInfo_t notifyTask;
         rtReduceAsyncTaskInfo_t reduceAsyncTask;
diff --git a/third_party/fwkacllib/inc/toolchain/prof_callback.h b/third_party/fwkacllib/inc/toolchain/prof_callback.h
index 18550157..5073cfb1 100644
--- a/third_party/fwkacllib/inc/toolchain/prof_callback.h
+++ b/third_party/fwkacllib/inc/toolchain/prof_callback.h
@@ -108,7 +108,19 @@ enum MsprofCtrlCallbackType {
     MSPROF_CTRL_INIT_ACL_ENV = 0,           // start profiling with acl env
     MSPROF_CTRL_INIT_ACL_JSON,              // start profiling with acl.json
     MSPROF_CTRL_INIT_GE_OPTIONS,            // start profiling with ge env and options
-    MSPROF_CTRL_FINALIZE                    // stop profiling
+    MSPROF_CTRL_FINALIZE,                   // stop profiling
+    MSPROF_CTRL_REPORT_FUN_P,               // for report callback
+    MSPROF_CTRL_PROF_SWITCH_ON,             // for prof switch on
+    MSPROF_CTRL_PROF_SWITCH_OFF             // for prof switch off
+};
+
+#define MSPROF_MAX_DEV_NUM (64)
+
+struct MsprofCommandHandle {
+    uint64_t profSwitch;
+    uint32_t devNums; // length of device id list
+    uint32_t devIdList[MSPROF_MAX_DEV_NUM];
+    uint32_t modelId;
 };
 
 /**
@@ -129,6 +141,23 @@ typedef int32_t (*MsprofCtrlCallback)(uint32_t type, void *data, uint32_t len);
  */
 typedef void (*MsprofSetDeviceCallback)(uint32_t devId, bool isOpenDevice);
 
+/*
+ * @name  MsprofInit
+ * @brief Profiling module init
+ * @param [in] dataType: profiling type: ACL Env/ACL Json/GE Option
+ * @param [in] data: profiling switch data
+ * @param [in] dataLen: Length of data
+ * @return 0:SUCCESS, >0:FAILED
+ */
+int32_t MsprofInit(uint32_t dataType, void *data, uint32_t dataLen);
+
+/*
+ * @name AscendCL
+ * @brief Finishing Profiling
+ * @param NULL
+ * @return 0:SUCCESS, >0:FAILED
+ */
+int32_t MsprofFinalize();
 #ifdef __cplusplus
 }
 #endif
diff --git a/third_party/fwkacllib/inc/toolchain/slog.h b/third_party/fwkacllib/inc/toolchain/slog.h
index ba286d02..cc7c83ca 100644
--- a/third_party/fwkacllib/inc/toolchain/slog.h
+++ b/third_party/fwkacllib/inc/toolchain/slog.h
@@ -17,6 +17,8 @@
 #ifndef D_SYSLOG_H_
 #define D_SYSLOG_H_
 
+static const int TMP_LOG = 0;
+
 #ifdef __cplusplus
 #ifndef LOG_CPP
 extern "C" {
@@ -120,15 +122,15 @@ typedef struct tagKV {
 } KeyValue;
 
 typedef enum {
-  APPLICATION = 0,
-  SYSTEM
+    APPLICATION = 0,
+    SYSTEM
 } ProcessType;
 
 typedef struct {
-  ProcessType type;
-  unsigned int pid;
-  unsigned int deviceId;
-  char reserved[RESERVERD_LENGTH];
+    ProcessType type;
+    unsigned int pid;
+    unsigned int deviceId;
+    char reserved[RESERVERD_LENGTH];
 } LogAttr;
 
 /**
@@ -141,7 +143,7 @@ enum {
   IDEDD,         /**< IDE daemon device */
   IDEDH,         /**< IDE daemon host */
   HCCL,          /**< HCCL */
-  FMK,           /**< Framework */
+  FMK,           /**< Adapter */
   HIAIENGINE,    /**< Matrix */
   DVPP,          /**< DVPP */
   RUNTIME,       /**< Runtime */
@@ -162,11 +164,11 @@ enum {
   MDCDEFAULT, /**< MDC undefine */
   MDCSC,      /**< MDC spatial cognition */
   MDCPNC,
-  MLL,
+  MLL,      /**< abandon */
   DEVMM,    /**< Dlog memory managent */
   KERNEL,   /**< Kernel */
   LIBMEDIA, /**< Libmedia */
-  CCECPU,   /**< ai cpu */
+  CCECPU,   /**< aicpu shedule */
   ASCENDDK, /**< AscendDK */
   ROS,      /**< ROS */
   HCCP,
@@ -179,7 +181,7 @@ enum {
   TSDUMP,    /**< TSDUMP module */
   AICPU,     /**< AICPU module */
   LP,        /**< LP module */
-  TDT,
+  TDT,       /**< tsdaemon or aicpu shedule */
   FE,
   MD,
   MB,
@@ -261,7 +263,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
 #define dlog_error(moduleId, fmt, ...)                                          \
   do {                                                                          \
     DlogErrorInner(moduleId, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__); \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -276,7 +278,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
     if(CheckLogLevel(moduleId, DLOG_WARN) == 1) {                                   \
         DlogWarnInner(moduleId, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__);  \
     }                                                                               \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -291,7 +293,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
     if(CheckLogLevel(moduleId, DLOG_INFO) == 1) {                                   \
         DlogInfoInner(moduleId, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__);  \
     }                                                                               \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -306,7 +308,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
     if(CheckLogLevel(moduleId, DLOG_DEBUG) == 1) {                                  \
         DlogDebugInner(moduleId, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__); \
     }                                                                               \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -318,7 +320,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
 #define dlog_event(moduleId, fmt, ...)                                          \
   do {                                                                          \
     DlogEventInner(moduleId, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__); \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -334,7 +336,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
     if(CheckLogLevel(moduleId, level) == 1) {                                           \
         DlogInner(moduleId, level, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__);   \
      }                                                                                  \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -351,7 +353,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
     if(CheckLogLevel(moduleId, level) == 1) {                                                           \
         DlogInner(moduleId, level, "[%s:%d][%s]" fmt, __FILE__, __LINE__, submodule, ##__VA_ARGS__);    \
     }                                                                                                   \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -369,7 +371,7 @@ DLL_EXPORT int DlogSetAttr(LogAttr logAttr);
     if(CheckLogLevel(moduleId, level) == 1) {                                                                   \
         DlogWithKVInner(moduleId, level, pstKVArray, kvNum, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__);  \
     }                                                                                                           \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -381,13 +383,13 @@ DLL_EXPORT void DlogFlush(void);
  * @ingroup slog
  * @brief Internal log interface, other modules are not allowed to call this interface
  */
-void DlogErrorInner(int moduleId, const char *fmt, ...) __attribute__((format(printf, 2, 3)));
-void DlogWarnInner(int moduleId, const char *fmt, ...) __attribute__((format(printf, 2, 3)));
-void DlogInfoInner(int moduleId, const char *fmt, ...) __attribute__((format(printf, 2, 3)));
-void DlogDebugInner(int moduleId, const char *fmt, ...) __attribute__((format(printf, 2, 3)));
-void DlogEventInner(int moduleId, const char *fmt, ...) __attribute__((format(printf, 2, 3)));
-void DlogInner(int moduleId, int level, const char *fmt, ...) __attribute__((format(printf, 3, 4)));
-void DlogWithKVInner(int moduleId, int level, KeyValue *pstKVArray, int kvNum, const char *fmt, ...) __attribute__((format(printf, 5, 6)));
+void DlogErrorInner(int moduleId, const char *fmt, ...);
+void DlogWarnInner(int moduleId, const char *fmt, ...);
+void DlogInfoInner(int moduleId, const char *fmt, ...);
+void DlogDebugInner(int moduleId, const char *fmt, ...);
+void DlogEventInner(int moduleId, const char *fmt, ...);
+void DlogInner(int moduleId, int level, const char *fmt, ...);
+void DlogWithKVInner(int moduleId, int level, KeyValue *pstKVArray, int kvNum, const char *fmt, ...);
 
 #ifdef __cplusplus
 #ifndef LOG_CPP
@@ -453,7 +455,7 @@ DLL_EXPORT int DlogSetAttrForC(LogAttr logAttr);
     if(CheckLogLevelForC(moduleId, level) == 1) {                                           \
         DlogInnerForC(moduleId, level, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__);   \
      }                                                                                  \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -470,7 +472,7 @@ DLL_EXPORT int DlogSetAttrForC(LogAttr logAttr);
     if(CheckLogLevelForC(moduleId, level) == 1) {                                                           \
         DlogInnerForC(moduleId, level, "[%s:%d][%s]" fmt, __FILE__, __LINE__, submodule, ##__VA_ARGS__);    \
     }                                                                                                   \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -488,7 +490,7 @@ DLL_EXPORT int DlogSetAttrForC(LogAttr logAttr);
     if(CheckLogLevelForC(moduleId, level) == 1) {                                                                   \
         DlogWithKVInnerForC(moduleId, level, pstKVArray, kvNum, "[%s:%d]" fmt, __FILE__, __LINE__, ##__VA_ARGS__);  \
     }                                                                                                           \
-  } while (0)
+  } while (TMP_LOG != 0)
 
 /**
  * @ingroup slog
@@ -500,8 +502,8 @@ DLL_EXPORT void DlogFlushForC(void);
  * @ingroup slog
  * @brief Internal log interface, other modules are not allowed to call this interface
  */
-void DlogInnerForC(int moduleId, int level, const char *fmt, ...) __attribute__((format(printf, 3, 4)));
-void DlogWithKVInnerForC(int moduleId, int level, KeyValue *pstKVArray, int kvNum, const char *fmt, ...) __attribute__((format(printf, 5, 6)));
+void DlogInnerForC(int moduleId, int level, const char *fmt, ...);
+void DlogWithKVInnerForC(int moduleId, int level, KeyValue *pstKVArray, int kvNum, const char *fmt, ...);
 
 #ifdef __cplusplus
 }
diff --git a/third_party/fwkacllib/inc/toolchain/tuning_tool/tune_api.h b/third_party/fwkacllib/inc/toolchain/tuning_tool/tune_api.h
index 6208f462..2cf6e0c4 100644
--- a/third_party/fwkacllib/inc/toolchain/tuning_tool/tune_api.h
+++ b/third_party/fwkacllib/inc/toolchain/tuning_tool/tune_api.h
@@ -1,72 +1,88 @@
-/**
- * @file tune_api.h
- *
- * Copyright (c) Huawei Technologies Co., Ltd. 2020-2020. All rights reserved.\n
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n
- * 描述：mstune调优接口头文件
- */
-/** @defgroup mstune mstune调优接口 */
-#ifndef TUNE_API_H
-#define TUNE_API_H
-#include <vector>
-#include <map>
-#include <string>
-#include "graph/graph.h"
-#include "ge/ge_api.h"
-
-/**
- * @ingroup mstune
- *
- * mstune status
- */
-enum MsTuneStatus {
-    MSTUNE_SUCCESS,  /** tune success */
-    MSTUNE_FAILED,   /** tune failed */
-};
-
-// Option key: for train options sets
-const std::string MSTUNE_SELF_KEY = "mstune";
-const std::string MSTUNE_GEINIT_KEY = "initialize";
-const std::string MSTUNE_GESESS_KEY = "session";
-
-/**
- * @ingroup mstune
- * @par 描述: 命令行调优
- *
- * @attention 无
- * @param  option [IN] 调优参数
- * @param  msg [OUT] 调优异常下返回信息
- * @retval #MSTUNE_SUCCESS 执行成功
- * @retval #MSTUNE_FAILED 执行失败
- * @par 依赖:
- * @li tune_api.cpp：该接口所属的开发包。
- * @li tune_api.h：该接口声明所在的头文件。
- * @see 无
- * @since
- */
-MsTuneStatus MsTuning(const std::map<std::string, std::string> &option, std::string &msg);
-
-/**
- * @ingroup mstune
- * @par 描述: 梯度调优
- *
- * @attention 无
- * @param  tuningGraph [IN] 调优图
- * @param  dependGraph [IN] 调优依赖图
- * @param  session [IN] ge连接会话
- * @param  option [IN] 参数集. 包含调优参数及ge参数
- * @retval #MSTUNE_SUCCESS 执行成功
- * @retval #MSTUNE_FAILED 执行失败
- * @par 依赖:
- * @li tune_api.cpp：该接口所属的开发包。
- * @li tune_api.h：该接口声明所在的头文件。
- * @see 无
- * @since
- */
-extern "C" MsTuneStatus MsTrainTuning(ge::Graph &tuningGraph, std::vector<ge::Graph> &dependGraph,
-    ge::Session *session, const std::map<std::string, std::map<std::string, std::string>> &option);
-
-#endif
+/**
+ * @file tune_api.h
+ *
+ * Copyright (c) Huawei Technologies Co., Ltd. 2020-2021. All rights reserved.\n
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n
+ * 描述：aoe调优接口头文件
+ */
+/** @defgroup aoe aoe调优接口 */
+#ifndef TUNE_API_H
+#define TUNE_API_H
+#include <map>
+#include <string>
+#include "ge/ge_api.h"
+#include "aoe_types.h"
+
+/**
+ * @ingroup aoe
+ * @par 描述: 命令行调优
+ *
+ * @attention 无
+ * @param  option [IN] 调优参数
+ * @param  msg [OUT] 调优异常下返回信息
+ * @retval #AOE_SUCCESS 执行成功
+ * @retval #AOE_FAILURE 执行失败
+ * @par 依赖:
+ * @li tune_api.cpp：该接口所属的开发包。
+ * @li tune_api.h：该接口声明所在的头文件。
+ * @see 无
+ * @since
+ */
+AoeStatus AoeOfflineTuning(const std::map<std::string, std::string> &option, std::string &msg);
+
+/**
+ * @ingroup aoe
+ * @par 描述: 调优初始化
+ *
+ * @attention 无
+ * @param  session [IN] ge连接会话
+ * @param  option [IN] 参数集. 包含调优参数及ge参数
+ * @retval #AOE_SUCCESS 执行成功
+ * @retval #AOE_FAILURE 执行失败
+ * @par 依赖:
+ * @li tune_api.cpp：该接口所属的开发包。
+ * @li tune_api.h：该接口声明所在的头文件。
+ * @see 无
+ * @since
+ */
+extern "C" AoeStatus AoeOnlineInitialize(ge::Session *session, const std::map<std::string, std::string> &option);
+
+/**
+ * @ingroup aoe
+ * @par 描述: 调优去初始化
+ *
+ * @attention 无
+ * @param  无
+ * @retval #AOE_SUCCESS 执行成功
+ * @retval #AOE_FAILURE 执行失败
+ * @par 依赖:
+ * @li tune_api.cpp：该接口所属的开发包。
+ * @li tune_api.h：该接口声明所在的头文件。
+ * @see 无
+ * @since
+ */
+extern "C" AoeStatus AoeOnlineFinalize();
+
+/**
+ * @ingroup aoe
+ * @par 描述: 调优处理
+ *
+ * @attention 无
+ * @param  tuningGraph [IN] 调优图
+ * @param  dependGraph [IN] 调优依赖图
+ * @param  session [IN] ge连接会话
+ * @param  option [IN] 参数集. 包含调优参数及ge参数
+ * @retval #AOE_SUCCESS 执行成功
+ * @retval #AOE_FAILURE 执行失败
+ * @par 依赖:
+ * @li tune_api.cpp：该接口所属的开发包。
+ * @li tune_api.h：该接口声明所在的头文件。
+ * @see 无
+ * @since
+ */
+extern "C" AoeStatus AoeOnlineTuning(ge::Graph &tuningGraph, std::vector<ge::Graph> &dependGraph,
+    ge::Session *session, const std::map<std::string, std::string> &option);
+#endif