Merge pull request #15646 from alalek:fix_avx512_detection

cmake/OpenCVCompilerOptimizations.cmake

@@ -5,13 +5,15 @@
 # AVX / AVX2 / AVX_512F
 # FMA3
 #
+# AVX512 details: https://en.wikipedia.org/wiki/AVX-512#CPUs_with_AVX-512
+#
 # CPU features groups:
 # AVX512_COMMON (Common instructions AVX-512F/CD for all CPUs that support AVX-512)
 # AVX512_KNL (Knights Landing with AVX-512F/CD/ER/PF)
 # AVX512_KNM (Knights Mill with AVX-512F/CD/ER/PF/4FMAPS/4VNNIW/VPOPCNTDQ)
 # AVX512_SKX (Skylake-X with AVX-512F/CD/BW/DQ/VL)
 # AVX512_CNL (Cannon Lake with AVX-512F/CD/BW/DQ/VL/IFMA/VBMI)
-# AVX512_CEL (Cascade Lake with AVX-512F/CD/BW/DQ/VL/IFMA/VBMI/VNNI)
+# AVX512_CLX (Cascade Lake with AVX-512F/CD/BW/DQ/VL/VNNI)
 # AVX512_ICL (Ice Lake with AVX-512F/CD/BW/DQ/VL/IFMA/VBMI/VNNI/VBMI2/BITALG/VPOPCNTDQ/VPCLMULQDQ*/GFNI*/VAES*)
 
 # ppc64le arch:
@@ -43,7 +45,7 @@
 # CPU_{opt}_ENABLED_DEFAULT=ON/OFF - has compiler support without additional flag (CPU_BASELINE_DETECT=ON only)
 
 set(CPU_ALL_OPTIMIZATIONS "SSE;SSE2;SSE3;SSSE3;SSE4_1;SSE4_2;POPCNT;AVX;FP16;AVX2;FMA3;AVX_512F")
-list(APPEND CPU_ALL_OPTIMIZATIONS "AVX512_COMMON;AVX512_KNL;AVX512_KNM;AVX512_SKX;AVX512_CNL;AVX512_CEL;AVX512_ICL")
+list(APPEND CPU_ALL_OPTIMIZATIONS "AVX512_COMMON;AVX512_KNL;AVX512_KNM;AVX512_SKX;AVX512_CNL;AVX512_CLX;AVX512_ICL")
 list(APPEND CPU_ALL_OPTIMIZATIONS NEON VFPV3 FP16)
 list(APPEND CPU_ALL_OPTIMIZATIONS MSA)
 list(APPEND CPU_ALL_OPTIMIZATIONS VSX VSX3)
@@ -163,15 +165,15 @@ elseif(" ${CMAKE_CXX_FLAGS} " MATCHES " -march=native | -xHost | /QxHost ")
 endif()
 
 if(X86 OR X86_64)
-  ocv_update(CPU_KNOWN_OPTIMIZATIONS "SSE;SSE2;SSE3;SSSE3;SSE4_1;POPCNT;SSE4_2;FP16;FMA3;AVX;AVX2;AVX_512F;AVX512_COMMON;AVX512_KNL;AVX512_KNM;AVX512_SKX;AVX512_CNL;AVX512_CEL;AVX512_ICL")
+  ocv_update(CPU_KNOWN_OPTIMIZATIONS "SSE;SSE2;SSE3;SSSE3;SSE4_1;POPCNT;SSE4_2;FP16;FMA3;AVX;AVX2;AVX_512F;AVX512_COMMON;AVX512_KNL;AVX512_KNM;AVX512_SKX;AVX512_CNL;AVX512_CLX;AVX512_ICL")
 
   ocv_update(CPU_AVX512_COMMON_GROUP "AVX_512F;AVX_512CD")
   ocv_update(CPU_AVX512_KNL_GROUP "AVX512_COMMON;AVX512_KNL_EXTRA")
   ocv_update(CPU_AVX512_KNM_GROUP "AVX512_KNL;AVX512_KNM_EXTRA;AVX_512VPOPCNTDQ")
   ocv_update(CPU_AVX512_SKX_GROUP "AVX512_COMMON;AVX_512VL;AVX_512BW;AVX_512DQ")
   ocv_update(CPU_AVX512_CNL_GROUP "AVX512_SKX;AVX_512IFMA;AVX_512VBMI")
-  ocv_update(CPU_AVX512_CEL_GROUP "AVX512_CNL;AVX_512VNNI")
-  ocv_update(CPU_AVX512_ICL_GROUP "AVX512_CEL;AVX_512VBMI2;AVX_512BITALG;AVX_512VPOPCNTDQ") # ? VPCLMULQDQ, GFNI, VAES
+  ocv_update(CPU_AVX512_CLX_GROUP "AVX512_SKX;AVX_512VNNI")
+  ocv_update(CPU_AVX512_ICL_GROUP "AVX512_SKX;AVX_512IFMA;AVX_512VBMI;AVX_512VNNI;AVX_512VBMI2;AVX_512BITALG;AVX_512VPOPCNTDQ") # ? VPCLMULQDQ, GFNI, VAES
 
   ocv_update(CPU_SSE_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_sse.cpp")
   ocv_update(CPU_SSE2_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_sse2.cpp")
@@ -189,12 +191,12 @@ if(X86 OR X86_64)
   ocv_update(CPU_AVX512_KNM_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_avx512knm.cpp")
   ocv_update(CPU_AVX512_SKX_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_avx512skx.cpp")
   ocv_update(CPU_AVX512_CNL_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_avx512cnl.cpp")
-  ocv_update(CPU_AVX512_CEL_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_avx512cel.cpp")
+  ocv_update(CPU_AVX512_CLX_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_avx512clx.cpp")
   ocv_update(CPU_AVX512_ICL_TEST_FILE "${OpenCV_SOURCE_DIR}/cmake/checks/cpu_avx512icl.cpp")
 
   if(NOT OPENCV_CPU_OPT_IMPLIES_IGNORE)
-    ocv_update(CPU_AVX512_ICL_IMPLIES "AVX512_CEL")
-    ocv_update(CPU_AVX512_CEL_IMPLIES "AVX512_CNL")
+    ocv_update(CPU_AVX512_ICL_IMPLIES "AVX512_SKX")
+    ocv_update(CPU_AVX512_CLX_IMPLIES "AVX512_SKX")
     ocv_update(CPU_AVX512_CNL_IMPLIES "AVX512_SKX")
     ocv_update(CPU_AVX512_SKX_IMPLIES "AVX512_COMMON")
     ocv_update(CPU_AVX512_KNM_IMPLIES "AVX512_KNL")
@@ -251,7 +253,7 @@ if(X86 OR X86_64)
     ocv_intel_compiler_optimization_option(AVX512_KNM "-xKNM" "/Qx:KNM")
     ocv_intel_compiler_optimization_option(AVX512_SKX "-xSKYLAKE-AVX512" "/Qx:SKYLAKE-AVX512")
     ocv_intel_compiler_optimization_option(AVX512_CNL "-xCANNONLAKE" "/Qx:CANNONLAKE")
-    ocv_intel_compiler_optimization_option(AVX512_CEL "-xCASCADELAKE" "/Qx:CASCADELAKE")
+    ocv_intel_compiler_optimization_option(AVX512_CLX "-xCASCADELAKE" "/Qx:CASCADELAKE")
     ocv_intel_compiler_optimization_option(AVX512_ICL "-xICELAKE-CLIENT" "/Qx:ICELAKE-CLIENT")
   elseif(CV_GCC OR CV_CLANG)
     ocv_update(CPU_AVX2_FLAGS_ON "-mavx2")

cmake/checks/cpu_avx512clx.cpp

@@ -3,9 +3,9 @@
 void test()
 {
     __m512i a, b, c;
-    a = _mm512_dpwssd_epi32(a, b, c);
+    a = _mm512_dpwssd_epi32(a, b, c);            // VNNI
 }
 #else
-#error "AVX512-CEL is not supported"
+#error "AVX512-CLX is not supported"
 #endif
-int main() { return 0; }
+int main() { return 0; }

cmake/checks/cpu_avx512icl.cpp

@@ -3,9 +3,10 @@
 void test()
 {
     __m512i a, b, c;
-    a = _mm512_popcnt_epi8(a);
-    a = _mm512_shrdv_epi64(a, b, c);
-    a = _mm512_popcnt_epi64(a);
+    a = _mm512_popcnt_epi8(a);                   // BITALG
+    a = _mm512_shrdv_epi64(a, b, c);             // VBMI2
+    a = _mm512_popcnt_epi64(a);                  // VPOPCNTDQ
+    a = _mm512_dpwssd_epi32(a, b, c);            // VNNI
 }
 #else
 #error "AVX512-ICL is not supported"

modules/core/include/opencv2/core/cv_cpu_dispatch.h

@@ -113,12 +113,18 @@
 #  define CV_AVX_512IFMA 1
 #  define CV_AVX_512VBMI 1
 #endif
-#ifdef CV_CPU_COMPILE_AVX512_CEL
-#  define CV_AVX512_CEL 1
+#ifdef CV_CPU_COMPILE_AVX512_CLX
+#  define CV_AVX512_CLX 1
 #  define CV_AVX_512VNNI 1
 #endif
 #ifdef CV_CPU_COMPILE_AVX512_ICL
 #  define CV_AVX512_ICL 1
+#  undef CV_AVX_512IFMA
+#  define CV_AVX_512IFMA 1
+#  undef CV_AVX_512VBMI
+#  define CV_AVX_512VBMI 1
+#  undef CV_AVX_512VNNI
+#  define CV_AVX_512VNNI 1
 #  define CV_AVX_512VBMI2 1
 #  define CV_AVX_512BITALG 1
 #  define CV_AVX_512VPOPCNTDQ 1
@@ -311,8 +317,8 @@ struct VZeroUpperGuard {
 #ifndef CV_AVX512_CNL
 #  define CV_AVX512_CNL 0
 #endif
-#ifndef CV_AVX512_CEL
-#  define CV_AVX512_CEL 0
+#ifndef CV_AVX512_CLX
+#  define CV_AVX512_CLX 0
 #endif
 #ifndef CV_AVX512_ICL
 #  define CV_AVX512_ICL 0

modules/core/include/opencv2/core/cv_cpu_helper.h

@@ -357,26 +357,26 @@
 #endif
 #define __CV_CPU_DISPATCH_CHAIN_AVX512_CNL(fn, args, mode, ...)  CV_CPU_CALL_AVX512_CNL(fn, args); __CV_EXPAND(__CV_CPU_DISPATCH_CHAIN_ ## mode(fn, args, __VA_ARGS__))
 
-#if !defined CV_DISABLE_OPTIMIZATION && defined CV_ENABLE_INTRINSICS && defined CV_CPU_COMPILE_AVX512_CEL
-#  define CV_TRY_AVX512_CEL 1
-#  define CV_CPU_FORCE_AVX512_CEL 1
-#  define CV_CPU_HAS_SUPPORT_AVX512_CEL 1
-#  define CV_CPU_CALL_AVX512_CEL(fn, args) return (cpu_baseline::fn args)
-#  define CV_CPU_CALL_AVX512_CEL_(fn, args) return (opt_AVX512_CEL::fn args)
-#elif !defined CV_DISABLE_OPTIMIZATION && defined CV_ENABLE_INTRINSICS && defined CV_CPU_DISPATCH_COMPILE_AVX512_CEL
-#  define CV_TRY_AVX512_CEL 1
-#  define CV_CPU_FORCE_AVX512_CEL 0
-#  define CV_CPU_HAS_SUPPORT_AVX512_CEL (cv::checkHardwareSupport(CV_CPU_AVX512_CEL))
-#  define CV_CPU_CALL_AVX512_CEL(fn, args) if (CV_CPU_HAS_SUPPORT_AVX512_CEL) return (opt_AVX512_CEL::fn args)
-#  define CV_CPU_CALL_AVX512_CEL_(fn, args) if (CV_CPU_HAS_SUPPORT_AVX512_CEL) return (opt_AVX512_CEL::fn args)
+#if !defined CV_DISABLE_OPTIMIZATION && defined CV_ENABLE_INTRINSICS && defined CV_CPU_COMPILE_AVX512_CLX
+#  define CV_TRY_AVX512_CLX 1
+#  define CV_CPU_FORCE_AVX512_CLX 1
+#  define CV_CPU_HAS_SUPPORT_AVX512_CLX 1
+#  define CV_CPU_CALL_AVX512_CLX(fn, args) return (cpu_baseline::fn args)
+#  define CV_CPU_CALL_AVX512_CLX_(fn, args) return (opt_AVX512_CLX::fn args)
+#elif !defined CV_DISABLE_OPTIMIZATION && defined CV_ENABLE_INTRINSICS && defined CV_CPU_DISPATCH_COMPILE_AVX512_CLX
+#  define CV_TRY_AVX512_CLX 1
+#  define CV_CPU_FORCE_AVX512_CLX 0
+#  define CV_CPU_HAS_SUPPORT_AVX512_CLX (cv::checkHardwareSupport(CV_CPU_AVX512_CLX))
+#  define CV_CPU_CALL_AVX512_CLX(fn, args) if (CV_CPU_HAS_SUPPORT_AVX512_CLX) return (opt_AVX512_CLX::fn args)
+#  define CV_CPU_CALL_AVX512_CLX_(fn, args) if (CV_CPU_HAS_SUPPORT_AVX512_CLX) return (opt_AVX512_CLX::fn args)
 #else
-#  define CV_TRY_AVX512_CEL 0
-#  define CV_CPU_FORCE_AVX512_CEL 0
-#  define CV_CPU_HAS_SUPPORT_AVX512_CEL 0
-#  define CV_CPU_CALL_AVX512_CEL(fn, args)
-#  define CV_CPU_CALL_AVX512_CEL_(fn, args)
+#  define CV_TRY_AVX512_CLX 0
+#  define CV_CPU_FORCE_AVX512_CLX 0
+#  define CV_CPU_HAS_SUPPORT_AVX512_CLX 0
+#  define CV_CPU_CALL_AVX512_CLX(fn, args)
+#  define CV_CPU_CALL_AVX512_CLX_(fn, args)
 #endif
-#define __CV_CPU_DISPATCH_CHAIN_AVX512_CEL(fn, args, mode, ...)  CV_CPU_CALL_AVX512_CEL(fn, args); __CV_EXPAND(__CV_CPU_DISPATCH_CHAIN_ ## mode(fn, args, __VA_ARGS__))
+#define __CV_CPU_DISPATCH_CHAIN_AVX512_CLX(fn, args, mode, ...)  CV_CPU_CALL_AVX512_CLX(fn, args); __CV_EXPAND(__CV_CPU_DISPATCH_CHAIN_ ## mode(fn, args, __VA_ARGS__))
 
 #if !defined CV_DISABLE_OPTIMIZATION && defined CV_ENABLE_INTRINSICS && defined CV_CPU_COMPILE_AVX512_ICL
 #  define CV_TRY_AVX512_ICL 1

modules/core/include/opencv2/core/cvdef.h

@@ -269,7 +269,7 @@ namespace cv { namespace debug_build_guard { } using namespace debug_build_guard
 #define CV_CPU_AVX512_KNL       258
 #define CV_CPU_AVX512_KNM       259
 #define CV_CPU_AVX512_CNL       260
-#define CV_CPU_AVX512_CEL       261
+#define CV_CPU_AVX512_CLX       261
 #define CV_CPU_AVX512_ICL       262
 
 // when adding to this list remember to update the following enum
@@ -320,7 +320,7 @@ enum CpuFeatures {
     CPU_AVX512_KNL      = 258, //!< Knights Landing with AVX-512F/CD/ER/PF
     CPU_AVX512_KNM      = 259, //!< Knights Mill with AVX-512F/CD/ER/PF/4FMAPS/4VNNIW/VPOPCNTDQ
     CPU_AVX512_CNL      = 260, //!< Cannon Lake with AVX-512F/CD/BW/DQ/VL/IFMA/VBMI
-    CPU_AVX512_CEL      = 261, //!< Cascade Lake with AVX-512F/CD/BW/DQ/VL/IFMA/VBMI/VNNI
+    CPU_AVX512_CLX      = 261, //!< Cascade Lake with AVX-512F/CD/BW/DQ/VL/VNNI
     CPU_AVX512_ICL      = 262, //!< Ice Lake with AVX-512F/CD/BW/DQ/VL/IFMA/VBMI/VNNI/VBMI2/BITALG/VPOPCNTDQ
 
     CPU_MAX_FEATURE     = 512  // see CV_HARDWARE_MAX_FEATURE

modules/core/src/system.cpp

@@ -370,11 +370,12 @@ struct HWFeatures
 
         g_hwFeatureNames[CPU_MSA] = "CPU_MSA";
 
+        g_hwFeatureNames[CPU_AVX512_COMMON] = "AVX512-COMMON";
         g_hwFeatureNames[CPU_AVX512_SKX] = "AVX512-SKX";
         g_hwFeatureNames[CPU_AVX512_KNL] = "AVX512-KNL";
         g_hwFeatureNames[CPU_AVX512_KNM] = "AVX512-KNM";
         g_hwFeatureNames[CPU_AVX512_CNL] = "AVX512-CNL";
-        g_hwFeatureNames[CPU_AVX512_CEL] = "AVX512-CEL";
+        g_hwFeatureNames[CPU_AVX512_CLX] = "AVX512-CLX";
         g_hwFeatureNames[CPU_AVX512_ICL] = "AVX512-ICL";
     }
 
@@ -485,9 +486,11 @@ struct HWFeatures
                                           have[CV_CPU_AVX_5124VNNIW] && have[CV_CPU_AVX_512VPOPCNTDQ];
                 have[CV_CPU_AVX512_SKX] = have[CV_CPU_AVX_512BW] && have[CV_CPU_AVX_512DQ] && have[CV_CPU_AVX_512VL];
                 have[CV_CPU_AVX512_CNL] = have[CV_CPU_AVX512_SKX] && have[CV_CPU_AVX_512IFMA] && have[CV_CPU_AVX_512VBMI];
-                have[CV_CPU_AVX512_CEL] = have[CV_CPU_AVX512_CNL] && have[CV_CPU_AVX_512VNNI];
-                have[CV_CPU_AVX512_ICL] = have[CV_CPU_AVX512_CEL] && have[CV_CPU_AVX_512VBMI2] &&
-                                          have[CV_CPU_AVX_512BITALG] && have[CV_CPU_AVX_512VPOPCNTDQ];
+                have[CV_CPU_AVX512_CLX] = have[CV_CPU_AVX512_SKX] && have[CV_CPU_AVX_512VNNI];
+                have[CV_CPU_AVX512_ICL] = have[CV_CPU_AVX512_SKX] &&
+                                          have[CV_CPU_AVX_512IFMA] && have[CV_CPU_AVX_512VBMI] &&
+                                          have[CV_CPU_AVX_512VNNI] &&
+                                          have[CV_CPU_AVX_512VBMI2] && have[CV_CPU_AVX_512BITALG] && have[CV_CPU_AVX_512VPOPCNTDQ];
             }
             else
             {
@@ -495,7 +498,7 @@ struct HWFeatures
                 have[CV_CPU_AVX512_KNM] = false;
                 have[CV_CPU_AVX512_SKX] = false;
                 have[CV_CPU_AVX512_CNL] = false;
-                have[CV_CPU_AVX512_CEL] = false;
+                have[CV_CPU_AVX512_CLX] = false;
                 have[CV_CPU_AVX512_ICL] = false;
             }
         }
@@ -572,8 +575,16 @@ struct HWFeatures
         have[CV_CPU_VSX3] = (CV_VSX3);
     #endif
 
+        bool skip_baseline_check = false;
+#ifndef NO_GETENV
+        if (getenv("OPENCV_SKIP_CPU_BASELINE_CHECK"))
+        {
+            skip_baseline_check = true;
+        }
+#endif
         int baseline_features[] = { CV_CPU_BASELINE_FEATURES };
-        if (!checkFeatures(baseline_features, sizeof(baseline_features) / sizeof(baseline_features[0])))
+        if (!checkFeatures(baseline_features, sizeof(baseline_features) / sizeof(baseline_features[0]))
+            && !skip_baseline_check)
         {
             fprintf(stderr, "\n"
                     "******************************************************************\n"
@@ -600,12 +611,12 @@ struct HWFeatures
             {
                 if (have[feature])
                 {
-                    if (dump) fprintf(stderr, "%s - OK\n", getHWFeatureNameSafe(feature));
+                    if (dump) fprintf(stderr, "    ID=%3d (%s) - OK\n", feature, getHWFeatureNameSafe(feature));
                 }
                 else
                 {
                     result = false;
-                    if (dump) fprintf(stderr, "%s - NOT AVAILABLE\n", getHWFeatureNameSafe(feature));
+                    if (dump) fprintf(stderr, "    ID=%3d (%s) - NOT AVAILABLE\n", feature, getHWFeatureNameSafe(feature));
                 }
             }
         }