Cycles: Increase minimum target on x86 to SSE4.2

* Compile regular host code with SSE4.2
* Remove the SSE2 kernel, only the SSE4.2 and AVX2 kernel remain

Pull Request: https://projects.blender.org/blender/blender/pulls/118471

intern/cycles/kernel/CMakeLists.txt

@@ -14,7 +14,6 @@ set(INC_SYS
 
 set(SRC_KERNEL_DEVICE_CPU
   device/cpu/kernel.cpp
-  device/cpu/kernel_sse2.cpp
   device/cpu/kernel_sse42.cpp
   device/cpu/kernel_avx2.cpp
 )
@@ -1163,8 +1162,7 @@ endif()
 
 set_source_files_properties(device/cpu/kernel.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_KERNEL_FLAGS}")
 
-if(CXX_HAS_SSE)
-  set_source_files_properties(device/cpu/kernel_sse2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE2_KERNEL_FLAGS}")
+if(CXX_HAS_SSE42)
   set_source_files_properties(device/cpu/kernel_sse42.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE42_KERNEL_FLAGS}")
 endif()
 

intern/cycles/kernel/device/cpu/kernel.cpp

@@ -4,31 +4,25 @@
 
 /* CPU kernel entry points */
 
-/* On x86-64, we can assume SSE2, so avoid the extra kernel and compile this
- * one with SSE2 intrinsics.
+/* On x86-64, our minimum is SSE4.2, so avoid the extra kernel and compile this
+ * one with SSE4.2 intrinsics.
  */
 #if defined(__x86_64__) || defined(_M_X64)
 #  define __KERNEL_SSE__
 #  define __KERNEL_SSE2__
+#  define __KERNEL_SSE3__
+#  define __KERNEL_SSSE3__
+#  define __KERNEL_SSE42__
 #endif
 
 /* When building kernel for native machine detect kernel features from the flags
  * set by compiler.
  */
 #ifdef WITH_KERNEL_NATIVE
-#  ifdef __SSE2__
-#    ifndef __KERNEL_SSE2__
-#      define __KERNEL_SSE2__
-#    endif
-#  endif
-#  ifdef __SSE3__
-#    define __KERNEL_SSE3__
-#  endif
-#  ifdef __SSSE3__
-#    define __KERNEL_SSSE3__
-#  endif
 #  ifdef __SSE4_2__
-#    define __KERNEL_SSE42__
+#    ifndef __KERNEL_SSE42__
+#      define __KERNEL_SSE42__
+#    endif
 #  endif
 #  ifdef __AVX__
 #    ifndef __KERNEL_SSE__

intern/cycles/kernel/device/cpu/kernel.h

@@ -33,9 +33,6 @@ void kernel_global_memory_copy(KernelGlobalsCPU *kg, const char *name, void *mem
 #define KERNEL_ARCH cpu
 #include "kernel/device/cpu/kernel_arch.h"
 
-#define KERNEL_ARCH cpu_sse2
-#include "kernel/device/cpu/kernel_arch.h"
-
 #define KERNEL_ARCH cpu_sse42
 #include "kernel/device/cpu/kernel_arch.h"
 

intern/cycles/kernel/device/cpu/kernel_sse2.cpp

@@ -1,22 +0,0 @@
-/* SPDX-FileCopyrightText: 2011-2022 Blender Foundation
- *
- * SPDX-License-Identifier: Apache-2.0 */
-
-/* Optimized CPU kernel entry points. This file is compiled with SSE2
- * optimization flags and nearly all functions inlined, while kernel.cpp
- * is compiled without for other CPU's. */
-
-#include "util/optimization.h"
-
-#ifndef WITH_CYCLES_OPTIMIZED_KERNEL_SSE2
-#  define KERNEL_STUB
-#else
-/* SSE optimization disabled for now on 32 bit, see bug #36316. */
-#  if !(defined(__GNUC__) && (defined(i386) || defined(_M_IX86)))
-#    define __KERNEL_SSE2__
-#  endif
-#endif /* WITH_CYCLES_OPTIMIZED_KERNEL_SSE2 */
-
-#include "kernel/device/cpu/kernel.h"
-#define KERNEL_ARCH cpu_sse2
-#include "kernel/device/cpu/kernel_arch_impl.h"

intern/cycles/kernel/device/cpu/kernel_sse42.cpp

@@ -2,7 +2,7 @@
  *
  * SPDX-License-Identifier: Apache-2.0 */
 
-/* Optimized CPU kernel entry points. This file is compiled with SSE3/SSSE3
+/* Optimized CPU kernel entry points. This file is compiled with SSE42
  * optimization flags and nearly all functions inlined, while kernel.cpp
  * is compiled without for other CPU's. */
 
@@ -13,6 +13,7 @@
 #else
 /* SSE optimization disabled for now on 32 bit, see bug #36316. */
 #  if !(defined(__GNUC__) && (defined(i386) || defined(_M_IX86)))
+#    define __KERNEL_SSE__
 #    define __KERNEL_SSE2__
 #    define __KERNEL_SSE3__
 #    define __KERNEL_SSSE3__