lowered the ISA requirement from SSE4.2 to SSE4.1

CHANGELOG.md

@@ -5,6 +5,7 @@ Version History
 
 -   Fixed wrong path in the generated CMake config
 -   Fixed wrong rpath in the binaries
+-   Lowered the CPU requirement to SSE4.1
 
 ### Changes in v0.8.0:
 

CMakeLists.txt

@@ -141,12 +141,12 @@ generate_cpp_resources(WEIGHTS_SOURCES "oidn::weights"
 
 add_library(${PROJECT_NAME} SHARED ${CORE_SOURCES} ${WEIGHTS_SOURCES})
 
-add_library("${PROJECT_NAME}_sse42" STATIC ${CORE_SOURCES_SSE42})
-set_target_properties("${PROJECT_NAME}_sse42" PROPERTIES COMPILE_FLAGS "${ISA_FLAGS_SSE42}")
+add_library("${PROJECT_NAME}_sse41" STATIC ${CORE_SOURCES_SSE42})
+set_target_properties("${PROJECT_NAME}_sse41" PROPERTIES COMPILE_FLAGS "${ISA_FLAGS_SSE41}")
 
 target_link_libraries(${PROJECT_NAME}
   PRIVATE
-    "${PROJECT_NAME}_sse42"
+    "${PROJECT_NAME}_sse41"
     common mkldnn
     ${TBB_LIBRARIES}
 )

core/device.cpp

@@ -23,8 +23,8 @@ namespace oidn {
 
   Device::Device()
   {
-    if (!mayiuse(sse42))
-      throw Exception(Error::UnsupportedHardware, "SSE4.2 support is required at minimum");
+    if (!mayiuse(sse41))
+      throw Exception(Error::UnsupportedHardware, "SSE4.1 support is required at minimum");
   }
 
   Device::~Device()

doc/api.md

@@ -99,7 +99,7 @@ can be one of the following:
 Name                     Description
 ------------------------ ------------------------------------------------------
 OIDN_DEVICE_TYPE_DEFAULT select the approximately fastest device
-OIDN_DEVICE_TYPE_CPU     CPU device (requires SSE4.2 support)
+OIDN_DEVICE_TYPE_CPU     CPU device (requires SSE4.1 support)
 ------------------------ ------------------------------------------------------
 : Supported device types, i.e., valid constants of type `OIDNDeviceType`.
 

doc/downloads.md

@@ -4,7 +4,7 @@ Download Precompiled Open Image Denoise Binary Packages
 Prerequisites
 -------------
 
-Your CPU must support at least SSE4.2 to run Open Image Denoise, and you need
+Your CPU must support at least SSE4.1 to run Open Image Denoise, and you need
 a 64-bit operating system as well. The TGZ/ZIP packages contain most needed
 3rd party dependencies.
 

doc/overview.md

@@ -24,7 +24,7 @@ ray tracing.
 Open Image Denoise internally builds on top of
 [Intel® Math Kernel Library for Deep Neural Networks (MKL-DNN)](https://github.com/intel/mkl-dnn),
 and fully exploits modern instruction sets like Intel SSE4, AVX2, and AVX-512
-to achieve high denoising performance. A CPU with support for at least SSE4.2
+to achieve high denoising performance. A CPU with support for at least SSE4.1
 is required to run Open Image Denoise.