If I read it correctly it says the floating point value does not fit in floating point type.
The supported compiler according to the release notes is
Intel® Parallel Studio XE for Windows* Version 2016 Update 3
You can also try building the project using the Visual Studio 2015 compiler.
I am not familiar enough with Intel System Studio to understand why there is that mismatch in floating point size. I would check compiler options to see what the size each component thinks should be.
This is a well known issue and it is because the SampleEnclave's project setting is configured to use the Intel C++ Compiler by default. Please either install Intel C++ compiler or change the project setting to use Visual Studio C++ Compiler. Please see attachment for an example on how to change the setting.
I have created a YouTube video for creating and compiling SGX application on the VS2015 from scratch. I have used Intel Parallel Studio XE 2017 along with SGX SDK 1.7 in the Video. Attached is the source codes that I have used in the video. I hope it helps.