diff --git a/cmake/cmake.define b/cmake/cmake.define
index 3343798686..56b6b7e1de 100644
--- a/cmake/cmake.define
+++ b/cmake/cmake.define
@@ -149,6 +149,8 @@ ELSE ()
     CHECK_C_COMPILER_FLAG("-mfma" COMPILER_SUPPORT_FMA)
     CHECK_C_COMPILER_FLAG("-mavx" COMPILER_SUPPORT_AVX)
     CHECK_C_COMPILER_FLAG("-mavx2" COMPILER_SUPPORT_AVX2)
+    CHECK_C_COMPILER_FLAG("-mavx512f" COMPILER_SUPPORT_AVX512F)
+    CHECK_C_COMPILER_FLAG("-mavx512vbmi" COMPILER_SUPPORT_AVX512BMI)
 
     IF (COMPILER_SUPPORT_SSE42)
         SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -msse4.2")
@@ -168,7 +170,13 @@ ELSE ()
             SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mavx2")
             SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mavx2")
         ENDIF()
-        MESSAGE(STATUS "SIMD instructions (FMA/AVX/AVX2) is ACTIVATED")
+        MESSAGE(STATUS "SIMD instructions (FMA/AVX/AVX2/AVX512) is ACTIVATED")
+
+        IF (COMPILER_SUPPORT_AVX512F AND COMPILER_SUPPORT_AVX512BMI)
+            SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mavx512f -mavx512vbmi")
+            SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mavx512f -mavx512vbmi")
+            MESSAGE(STATUS "avx512 supported by gcc")
+        ENDIF()
     ENDIF()
 
     # build mode
diff --git a/include/os/osEnv.h b/include/os/osEnv.h
index bc65da47a9..ac4ecd4212 100644
--- a/include/os/osEnv.h
+++ b/include/os/osEnv.h
@@ -36,11 +36,12 @@ extern int64_t         tsStreamMax;
 extern float           tsNumOfCores;
 extern int64_t         tsTotalMemoryKB;
 extern char           *tsProcPath;
-extern char            tsSIMDBuiltins;
+extern char            tsSIMDEnable;
 extern char            tsSSE42Enable;
 extern char            tsAVXEnable;
 extern char            tsAVX2Enable;
 extern char            tsFMAEnable;
+extern char 	       tsAVX512Enable;
 extern char            tsTagFilterCache;
 
 extern char configDir[];
diff --git a/include/os/osSysinfo.h b/include/os/osSysinfo.h
index 29b6f07dca..7a1df2b81c 100644
--- a/include/os/osSysinfo.h
+++ b/include/os/osSysinfo.h
@@ -41,7 +41,7 @@ int32_t taosGetOsReleaseName(char *releaseName, char* sName, char* ver, int32_t
 int32_t taosGetCpuInfo(char *cpuModel, int32_t maxLen, float *numOfCores);
 int32_t taosGetCpuCores(float *numOfCores, bool physical);
 void    taosGetCpuUsage(double *cpu_system, double *cpu_engine);
-int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma);
+int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma, char* avx512);
 int32_t taosGetTotalMemory(int64_t *totalKB);
 int32_t taosGetProcMemory(int64_t *usedKB);
 int32_t taosGetSysMemory(int64_t *usedKB);
diff --git a/source/common/src/tglobal.c b/source/common/src/tglobal.c
index cb67fc1ba3..d12ebb13c2 100644
--- a/source/common/src/tglobal.c
+++ b/source/common/src/tglobal.c
@@ -511,12 +511,13 @@ static int32_t taosAddSystemCfg(SConfig *pCfg) {
   if (cfgAddBool(pCfg, "enableCoreFile", 1, CFG_SCOPE_BOTH, CFG_DYN_CLIENT) != 0) return -1;
   if (cfgAddFloat(pCfg, "numOfCores", tsNumOfCores, 1, 100000, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
 
-  if (cfgAddBool(pCfg, "ssd42", tsSSE42Enable, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
-  if (cfgAddBool(pCfg, "avx", tsAVXEnable, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
-  if (cfgAddBool(pCfg, "avx2", tsAVX2Enable, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
-  if (cfgAddBool(pCfg, "fma", tsFMAEnable, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
-  if (cfgAddBool(pCfg, "simdEnable", tsSIMDBuiltins, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
-  if (cfgAddBool(pCfg, "tagFilterCache", tsTagFilterCache, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
+  if (cfgAddBool(pCfg, "ssd42", tsSSE42Enable, CFG_SCOPE_BOTH) != 0) return -1;
+  if (cfgAddBool(pCfg, "avx", tsAVXEnable, CFG_SCOPE_BOTH) != 0) return -1;
+  if (cfgAddBool(pCfg, "avx2", tsAVX2Enable, CFG_SCOPE_BOTH) != 0) return -1;
+  if (cfgAddBool(pCfg, "fma", tsFMAEnable, CFG_SCOPE_BOTH) != 0) return -1;
+  if (cfgAddBool(pCfg, "avx512", tsAVX512Enable, CFG_SCOPE_BOTH) != 0) return -1;
+  if (cfgAddBool(pCfg, "simdEnable", tsSIMDEnable, CFG_SCOPE_BOTH) != 0) return -1;
+  if (cfgAddBool(pCfg, "tagFilterCache", tsTagFilterCache, CFG_SCOPE_BOTH) != 0) return -1;
 
   if (cfgAddInt64(pCfg, "openMax", tsOpenMax, 0, INT64_MAX, CFG_SCOPE_BOTH, CFG_DYN_NONE) != 0) return -1;
 #if !defined(_ALPINE)
@@ -1080,7 +1081,7 @@ static int32_t taosSetServerCfg(SConfig *pCfg) {
   tsNumOfSnodeWriteThreads = cfgGetItem(pCfg, "numOfSnodeUniqueThreads")->i32;
   tsRpcQueueMemoryAllowed = cfgGetItem(pCfg, "rpcQueueMemoryAllowed")->i64;
 
-  tsSIMDBuiltins = (bool)cfgGetItem(pCfg, "simdEnable")->bval;
+  tsSIMDEnable = (bool)cfgGetItem(pCfg, "simdEnable")->bval;
   tsTagFilterCache = (bool)cfgGetItem(pCfg, "tagFilterCache")->bval;
 
   tsEnableMonitor = cfgGetItem(pCfg, "monitor")->bval;
diff --git a/source/libs/function/src/detail/tavgfunction.c b/source/libs/function/src/detail/tavgfunction.c
index 50df1b5067..e626c937da 100644
--- a/source/libs/function/src/detail/tavgfunction.c
+++ b/source/libs/function/src/detail/tavgfunction.c
@@ -565,7 +565,7 @@ int32_t avgFunction(SqlFunctionCtx* pCtx) {
     numOfElem = pInput->numOfRows;
     pAvgRes->count += pInput->numOfRows;
 
-    bool simdAvailable = tsAVXEnable && tsSIMDBuiltins && (numOfRows > THRESHOLD_SIZE);
+    bool simdAvailable = tsAVXEnable && tsSIMDEnable && (numOfRows > THRESHOLD_SIZE);
 
     switch(type) {
       case TSDB_DATA_TYPE_UTINYINT:
diff --git a/source/libs/function/src/detail/tminmax.c b/source/libs/function/src/detail/tminmax.c
index 3ca1c06303..a6c91a57ce 100644
--- a/source/libs/function/src/detail/tminmax.c
+++ b/source/libs/function/src/detail/tminmax.c
@@ -370,7 +370,7 @@ static int32_t findFirstValPosition(const SColumnInfoData* pCol, int32_t start,
 static void handleInt8Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc,
                           bool signVal) {
   // AVX2 version to speedup the loop
-  if (tsAVX2Enable && tsSIMDBuiltins) {
+  if (tsAVX2Enable && tsSIMDEnable) {
     pBuf->v = i8VectorCmpAVX2(data, numOfRows, isMinFunc, signVal);
   } else {
     if (!pBuf->assign) {
@@ -404,7 +404,7 @@ static void handleInt8Col(const void* data, int32_t start, int32_t numOfRows, SM
 static void handleInt16Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc,
                            bool signVal) {
   // AVX2 version to speedup the loop
-  if (tsAVX2Enable && tsSIMDBuiltins) {
+  if (tsAVX2Enable && tsSIMDEnable) {
     pBuf->v = i16VectorCmpAVX2(data, numOfRows, isMinFunc, signVal);
   } else {
     if (!pBuf->assign) {
@@ -438,7 +438,7 @@ static void handleInt16Col(const void* data, int32_t start, int32_t numOfRows, S
 static void handleInt32Col(const void* data, int32_t start, int32_t numOfRows, SMinmaxResInfo* pBuf, bool isMinFunc,
                            bool signVal) {
   // AVX2 version to speedup the loop
-  if (tsAVX2Enable && tsSIMDBuiltins) {
+  if (tsAVX2Enable && tsSIMDEnable) {
     pBuf->v = i32VectorCmpAVX2(data, numOfRows, isMinFunc, signVal);
   } else {
     if (!pBuf->assign) {
@@ -502,7 +502,7 @@ static void handleFloatCol(SColumnInfoData* pCol, int32_t start, int32_t numOfRo
   float* val = (float*)&pBuf->v;
 
   // AVX version to speedup the loop
-  if (tsAVXEnable && tsSIMDBuiltins) {
+  if (tsAVXEnable && tsSIMDEnable) {
     *val = floatVectorCmpAVX(pData, numOfRows, isMinFunc);
   } else {
     if (!pBuf->assign) {
@@ -533,7 +533,7 @@ static void handleDoubleCol(SColumnInfoData* pCol, int32_t start, int32_t numOfR
   double* val = (double*)&pBuf->v;
 
   // AVX version to speedup the loop
-  if (tsAVXEnable && tsSIMDBuiltins) {
+  if (tsAVXEnable && tsSIMDEnable) {
     *val = (double)doubleVectorCmpAVX(pData, numOfRows, isMinFunc);
   } else {
     if (!pBuf->assign) {
diff --git a/source/os/src/osEnv.c b/source/os/src/osEnv.c
index 0fc136c693..54107db325 100644
--- a/source/os/src/osEnv.c
+++ b/source/os/src/osEnv.c
@@ -37,11 +37,12 @@ float           tsNumOfCores = 0;
 int64_t         tsTotalMemoryKB = 0;
 char           *tsProcPath = NULL;
 
-char tsSIMDBuiltins = 0;
+char tsSIMDEnable = 0;
 char tsSSE42Enable = 0;
 char tsAVXEnable = 0;
 char tsAVX2Enable = 0;
 char tsFMAEnable = 0;
+char tsAVX512Enable = 0;
 
 void osDefaultInit() {
   taosSeedRand(taosSafeRand());
diff --git a/source/os/src/osSysinfo.c b/source/os/src/osSysinfo.c
index 4816ec8f8b..fea7a4f63d 100644
--- a/source/os/src/osSysinfo.c
+++ b/source/os/src/osSysinfo.c
@@ -250,7 +250,7 @@ void taosGetSystemInfo() {
   taosGetCpuCores(&tsNumOfCores, false);
   taosGetTotalMemory(&tsTotalMemoryKB);
   taosGetCpuUsage(NULL, NULL);
-  taosGetCpuInstructions(&tsSSE42Enable, &tsAVXEnable, &tsAVX2Enable, &tsFMAEnable);
+  taosGetCpuInstructions(&tsSSE42Enable, &tsAVXEnable, &tsAVX2Enable, &tsFMAEnable, &tsAVX512Enable);
 #endif
 }
 
@@ -602,7 +602,7 @@ void taosGetCpuUsage(double *cpu_system, double *cpu_engine) {
                       : "0"(level))
 
 // todo add for windows and mac
-int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma) {
+int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma, char* avx512) {
 #ifdef WINDOWS
 #elif defined(_TD_DARWIN_64)
 #else
@@ -610,12 +610,6 @@ int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma) {
 #ifdef _TD_X86_
   // Since the compiler is not support avx/avx2 instructions, the global variables always need to be
   // set to be false
-//#if __AVX__ || __AVX2__
-//  tsSIMDBuiltins = true;
-//#else
-//  tsSIMDBuiltins = false;
-//#endif
-
   uint32_t eax = 0, ebx = 0, ecx = 0, edx = 0;
 
   int32_t ret = __get_cpuid(1, &eax, &ebx, &ecx, &edx);
@@ -631,6 +625,7 @@ int32_t taosGetCpuInstructions(char* sse42, char* avx, char* avx2, char* fma) {
   // Ref to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77756
   __cpuid_fix(7u, eax, ebx, ecx, edx);
   *avx2 = (char) ((ebx & bit_AVX2) == bit_AVX2);
+  *avx512 = (char)((ebx & bit_AVX512F) == bit_AVX512F);
 #endif   // _TD_X86_
 #endif
 
diff --git a/source/util/src/tcompression.c b/source/util/src/tcompression.c
index 3fc3ef6be6..dc89a24180 100644
--- a/source/util/src/tcompression.c
+++ b/source/util/src/tcompression.c
@@ -283,7 +283,7 @@ int32_t tsDecompressINTImp(const char *const input, const int32_t nelements, cha
         int32_t batch = num >> 2;
         int32_t remain = num & 0x03;
         if (selector == 0 || selector == 1) {
-          if (tsAVX2Enable && tsSIMDBuiltins) {
+          if (tsAVX2Enable && tsSIMDEnable) {
             for (int32_t i = 0; i < batch; ++i) {
               __m256i prev = _mm256_set1_epi64x(prev_value);
               _mm256_storeu_si256((__m256i *)&p[_pos], prev);
@@ -300,7 +300,7 @@ int32_t tsDecompressINTImp(const char *const input, const int32_t nelements, cha
             }
           }
         } else {
-          if (tsAVX2Enable && tsSIMDBuiltins) {
+          if (tsAVX2Enable && tsSIMDEnable) {
             __m256i base = _mm256_set1_epi64x(w);
             __m256i maskVal = _mm256_set1_epi64x(mask);