mirror of
https://github.com/opencv/opencv.git
synced 2026-07-31 00:03:03 +04:00
fix avx512 detection
- renamed Cascade Lake AVX512_CEL => AVX512_CLX (align with Intel SDE tool) - fixed CLX instruction sets (no IFMA/VBMI) - added flag to bypass CPU baseline check: OPENCV_SKIP_CPU_BASELINE_CHECK
This commit is contained in:
@@ -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
|
||||
|
||||
@@ -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
|
||||
|
||||
@@ -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
|
||||
|
||||
@@ -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));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user