sln: remove SSE4 from build options, make SSE4 default

PCSX2 · Mar 26, 2021 · 8b864e1 · 8b864e1
1 parent 1c88a2d
commit 8b864e1
Show file tree

Hide file tree

Showing 11 changed files with 109 additions and 192 deletions.
diff --git a/PCSX2_suite.sln b/PCSX2_suite.sln
diff --git a/buildbot.xml b/buildbot.xml
@@ -53,7 +53,7 @@
     </PropertyGroup>
     <ItemGroup>
       <ConfigCPU Include="Debug"/>
-      <!--ConfigCPU Include="Debug AVX2;Debug SSE4"/-->
+      <!--ConfigCPU Include="Debug AVX2;Debug"/-->
     </ItemGroup>
   </Target>
   <Target Name="DevelAll">
@@ -62,15 +62,15 @@
     </PropertyGroup>
     <ItemGroup>
       <ConfigCPU Include="Devel"/>
-      <!--ConfigCPU Include="Devel AVX2;Devel SSE4"/-->
+      <!--ConfigCPU Include="Devel AVX2;Devel"/-->
     </ItemGroup>
   </Target>
   <Target Name="ReleaseAll">
     <PropertyGroup>
       <BaseConfiguration>Release</BaseConfiguration>
     </PropertyGroup>
     <ItemGroup>
-      <ConfigCPU Include="Release AVX2;Release SSE4"/>
+      <ConfigCPU Include="Release AVX2;Release"/>
     </ItemGroup>
   </Target>
 </Project>
diff --git a/plugins/GSdx/CMakeLists.txt b/plugins/GSdx/CMakeLists.txt
@@ -270,9 +270,7 @@ if(BUILTIN_GS)
 else()
     if (DISABLE_ADVANCE_SIMD)
         # Don't append -SSE2 on the first build to keep same name as SIMD build
-        add_pcsx2_plugin("${Output}" "${GSdxFinalSources}" "${GSdxFinalLibs}" "${GSdxFinalFlags}")
-        add_pcsx2_plugin("${Output}-SSE4" "${GSdxFinalSources}" "${GSdxFinalLibs}" "${GSdxFinalFlags} -mssse3 -msse4 -msse4.1")
-        target_compile_features("${Output}-SSE4" PRIVATE cxx_std_17)
+        add_pcsx2_plugin("${Output}" "${GSdxFinalSources}" "${GSdxFinalLibs}" "${GSdxFinalFlags} -msse3 -msse4 -msse4.1")
         add_pcsx2_plugin("${Output}-AVX2" "${GSdxFinalSources}" "${GSdxFinalLibs}" "${GSdxFinalFlags} -mavx -mavx2 -mbmi -mbmi2")
         target_compile_features("${Output}-AVX2" PRIVATE cxx_std_17)
     else()

diff --git a/plugins/GSdx/GSBlock.h b/plugins/GSdx/GSBlock.h
@@ -519,6 +519,35 @@ class GSBlock
 
 	template<int i> __forceinline static void ReadColumn8(const uint8* RESTRICT src, uint8* RESTRICT dst, int dstpitch)
 	{
+
+		//for(int j = 0; j < 64; j++) ((uint8*)src)[j] = (uint8)j;
+
+		#if 0 //_M_SSE >= 0x501
+
+		const GSVector8i* s = (const GSVector8i*)src;
+
+		GSVector8i v0 = s[i * 2 + 0];
+		GSVector8i v1 = s[i * 2 + 1];
+
+		GSVector8i::sw8(v0, v1);
+		GSVector8i::sw16(v0, v1);
+		GSVector8i::sw8(v0, v1);
+		GSVector8i::sw128(v0, v1);
+		GSVector8i::sw16(v0, v1);
+
+		v0 = v0.acbd();
+		v1 = v1.acbd();
+		v1 = v1.yxwz();
+
+		GSVector8i::storel(&dst[dstpitch * 0], v0);
+		GSVector8i::storeh(&dst[dstpitch * 1], v0);
+		GSVector8i::storel(&dst[dstpitch * 2], v1);
+		GSVector8i::storeh(&dst[dstpitch * 3], v1);
+
+		// TODO: not sure if this is worth it, not in this form, there should be a shorter path
+
+		#else
+
 		const GSVector4i* s = (const GSVector4i*)src;
 
 		GSVector4i v0, v1, v2, v3;
@@ -550,6 +579,8 @@ class GSBlock
 		GSVector4i::store<true>(&dst[dstpitch * 1], v3);
 		GSVector4i::store<true>(&dst[dstpitch * 2], v1);
 		GSVector4i::store<true>(&dst[dstpitch * 3], v2);
+
+		#endif
 	}
 
 	template<int i> __forceinline static void ReadColumn4(const uint8* RESTRICT src, uint8* RESTRICT dst, int dstpitch)

diff --git a/plugins/GSdx/GSState.cpp b/plugins/GSdx/GSState.cpp
@@ -2734,11 +2734,7 @@ __forceinline void GSState::VertexKick(uint32 skip)
 
 	GSVector4i xy = v1.xxxx().u16to32().sub32(m_ofxy);
 
-#if _M_SSE >= 0x401
 	GSVector4i::storel(&m_vertex.xy[xy_tail & 3], xy.blend16<0xf0>(xy.sra32(4)).ps32());
-#else
-	GSVector4i::storel(&m_vertex.xy[xy_tail & 3], xy.upl64(xy.sra32(4).zwzw()).ps32());
-#endif
 
 	m_vertex.tail = ++tail;
 	m_vertex.xy_tail = ++xy_tail;

diff --git a/plugins/GSdx/GSUtil.cpp b/plugins/GSdx/GSUtil.cpp
@@ -85,10 +85,6 @@ const char* GSUtil::GetLibName()
 		"AVX", sw_sse
 #elif _M_SSE >= 0x401
 		"SSE4.1", sw_sse
-#elif _M_SSE >= 0x301
-		"SSSE3", sw_sse
-#elif _M_SSE >= 0x200
-		"SSE2", sw_sse
 #endif
 	);
 
@@ -221,13 +217,7 @@ bool GSUtil::CheckSSE()
 	};
 
 	ISA checks[] = {
-		{Xbyak::util::Cpu::tSSE2, "SSE2"},
-#if _M_SSE >= 0x301
-		{Xbyak::util::Cpu::tSSSE3, "SSSE3"},
-#endif
-#if _M_SSE >= 0x401
 		{Xbyak::util::Cpu::tSSE41, "SSE41"},
-#endif
 #if _M_SSE >= 0x500
 		{Xbyak::util::Cpu::tAVX, "AVX1"},
 #endif

diff --git a/plugins/GSdx/GSVector4.h b/plugins/GSdx/GSVector4.h
@@ -316,22 +316,54 @@ class alignas(16) GSVector4
 
 	__forceinline GSVector4 madd(const GSVector4& a, const GSVector4& b) const
 	{
-		return *this * a + b;
+		#if 0//_M_SSE >= 0x501
+
+		return GSVector4(_mm_fmadd_ps(m, a, b));
+
+		#else
+
+ 		return *this * a + b;
+
+		#endif
 	}
 
 	__forceinline GSVector4 msub(const GSVector4& a, const GSVector4& b) const
 	{
-		return *this * a - b;
+		#if 0//_M_SSE >= 0x501
+
+		return GSVector4(_mm_fmsub_ps(m, a, b));
+
+		#else
+
+ 		return *this * a - b;
+
+		#endif
 	}
 
 	__forceinline GSVector4 nmadd(const GSVector4& a, const GSVector4& b) const
 	{
-		return b - *this * a;
+		#if 0//_M_SSE >= 0x501
+
+		return GSVector4(_mm_fnmadd_ps(m, a, b));
+
+		#else
+
+ 		return b - *this * a;
+
+		#endif	
 	}
 
 	__forceinline GSVector4 nmsub(const GSVector4& a, const GSVector4& b) const
 	{
-		return -b - *this * a;
+		#if 0//_M_SSE >= 0x501
+
+		return GSVector4(_mm_fnmsub_ps(m, a, b));
+
+		#else
+
+ 		return -b - *this * a;
+
+		#endif
 	}
 
 	__forceinline GSVector4 addm(const GSVector4& a, const GSVector4& b) const
@@ -460,16 +492,10 @@ class alignas(16) GSVector4
 
 		return _mm_testz_ps(m, m) != 0;
 
-		#elif _M_SSE >= 0x401
-
 		__m128i a = _mm_castps_si128(m);
 
 		return _mm_testz_si128(a, a) != 0;
 
-		#else
-
-		return mask() == 0;
-
 		#endif
 	}
 
@@ -482,6 +508,15 @@ class alignas(16) GSVector4
 	{
 		// TODO: use blendps when src == dst
 
+
+		#if 0 // _M_SSE >= 0x401
+
+		// NOTE: it's faster with shuffles...
+
+		return GSVector4(_mm_insert_ps(m, v.m, _MM_MK_INSERTPS_NDX(src, dst, 0)));
+
+		#else
+
 		switch(dst)
 		{
 		case 0:
@@ -527,6 +562,8 @@ class alignas(16) GSVector4
 		default:
 			__assume(0);
 		}
+
+		#endif
 	}
 
 #ifdef __linux__

diff --git a/plugins/GSdx/GSVector4i.h b/plugins/GSdx/GSVector4i.h
@@ -533,7 +533,15 @@ class alignas(16) GSVector4i
 
 	__forceinline GSVector4i upl8() const
 	{
-		return GSVector4i(_mm_unpacklo_epi8(m, _mm_setzero_si128()));
+		#if 0 // _M_SSE >= 0x401 // TODO: compiler bug
+
+		return GSVector4i(_mm_cvtepu8_epi16(m));
+
+		#else
+
+ 		return GSVector4i(_mm_unpacklo_epi8(m, _mm_setzero_si128()));
+
+		#endif	
 	}
 
 	__forceinline GSVector4i uph8() const
@@ -543,7 +551,15 @@ class alignas(16) GSVector4i
 
 	__forceinline GSVector4i upl16() const
 	{
-		return GSVector4i(_mm_unpacklo_epi16(m, _mm_setzero_si128()));
+		#if 0 //_M_SSE >= 0x401 // TODO: compiler bug
+
+		return GSVector4i(_mm_cvtepu16_epi32(m));
+
+		#else
+
+ 		return GSVector4i(_mm_unpacklo_epi16(m, _mm_setzero_si128()));
+
+		#endif	
 	}
 
 	__forceinline GSVector4i uph16() const
@@ -553,7 +569,15 @@ class alignas(16) GSVector4i
 
 	__forceinline GSVector4i upl32() const
 	{
-		return GSVector4i(_mm_unpacklo_epi32(m, _mm_setzero_si128()));
+		#if 0 //_M_SSE >= 0x401 // TODO: compiler bug
+
+		return GSVector4i(_mm_cvtepu32_epi64(m));
+
+		#else
+
+ 		return GSVector4i(_mm_unpacklo_epi32(m, _mm_setzero_si128()));
+
+		#endif	
 	}
 
 	__forceinline GSVector4i uph32() const