diff --git a/Ratatouille/Ratatouille.cpp b/Ratatouille/Ratatouille.cpp index 7443ded..1083329 100644 --- a/Ratatouille/Ratatouille.cpp +++ b/Ratatouille/Ratatouille.cpp @@ -61,7 +61,7 @@ namespace ratatouille { class DenormalProtection { private: -#ifdef __SSE__ +#ifdef USE_SSE uint32_t mxcsr_mask; uint32_t mxcsr; uint32_t old_mxcsr; @@ -69,20 +69,20 @@ class DenormalProtection { public: inline void set_() { -#ifdef __SSE__ +#ifdef USE_SSE old_mxcsr = _mm_getcsr(); mxcsr = old_mxcsr; _mm_setcsr((mxcsr | _MM_DENORMALS_ZERO_MASK | _MM_FLUSH_ZERO_MASK) & mxcsr_mask); #endif }; inline void reset_() { -#ifdef __SSE__ +#ifdef USE_SSE _mm_setcsr(old_mxcsr); #endif }; inline DenormalProtection() { -#ifdef __SSE__ +#ifdef USE_SSE mxcsr_mask = 0xffbf; // Default MXCSR mask mxcsr = 0; uint8_t fxsave[512] __attribute__ ((aligned (16))); // Structure for storing FPU state with FXSAVE command diff --git a/Ratatouille/makefile b/Ratatouille/makefile index 7072b1e..3ef2c04 100644 --- a/Ratatouille/makefile +++ b/Ratatouille/makefile @@ -45,13 +45,13 @@ endif # check for sse optimisation level only on x86_64 architecture ifeq ($(TARGET_ARCH), x86_64) ifneq ($$(filter $(CPU_INFO) | grep sse3 ) , ) - SSE_CFLAGS = -msse3 -mfpmath=sse -mfxsr + SSE_CFLAGS = -msse3 -mfpmath=sse -mfxsr -DUSE_SSE FFT_FLAG = -DFFTCONVOLVER_USE_SSE=1 else ifneq ($$(filter $(CPU_INFO) | grep sse2 ) , ) - SSE_CFLAGS = -msse2 -mfpmath=sse -mfxsr + SSE_CFLAGS = -msse2 -mfpmath=sse -mfxsr -DUSE_SSE FFT_FLAG = -DFFTCONVOLVER_USE_SSE=1 else ifneq ($$(filter $(CPU_INFO) | grep sse ) , ) - SSE_CFLAGS = -msse -mfpmath=sse -mfxsr + SSE_CFLAGS = -msse -mfpmath=sse -mfxsr -DUSE_SSE FFT_FLAG = -DFFTCONVOLVER_USE_SSE=1 else ifneq ($$(filter $(CPU_INFO) | grep ARM ) , ) ifneq ($$(filter $(CPU_INFO) | grep ARMv7 ) , ) @@ -99,15 +99,15 @@ CXX_v = $(shell $(CXX) -dumpversion) ifneq ($(MAKECMDGOALS),install) ifneq ($(MAKECMDGOALS),clean) ifneq ($(NO_x86_64), $(x86_64_LEVEL)) - ifeq ($(shell awk -v a="$(CXX_v)" -v b="11" 'BEGIN{print(a