Open prehensilecode opened 5 years ago
You can put whatever you want into
make/local
but icpc isn't a big priority for us.
On Fri, Mar 15, 2019 at 12:09 PM David Chin notifications@github.com wrote:
Description
Intel C++ compiler (icpc) should be supported, together with the "-xHost" compiler option to generate instructions for the highest instruction set and processor available on the compilation host machine. Example
In make/default_compiler_options set
CC=icpc
Currently, since the optimization flag only sets the integer optimization level, there is no mechanism for adding other optimization options (e.g. for code generation). Expected Output
Example with Intel Composer XE 2019u1:
Stan Math makefile: Current configuration:
- OS_TYPE (Operating System): linux
- CC (Compiler): icpc
- CC_TYPE intel
- Compiler version: 19.0.1.144 20181018
- O (Optimization Level): 3
Current Version:
v2.18.1
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/stan-dev/math/issues/1160, or mute the thread https://github.com/notifications/unsubscribe-auth/ADOrql0cywo_X_bywJ7czCxBinOy-ZSPks5vW8W5gaJpZM4b2z06 .
We'll also accept pull requests to bring our math lib up to other library standards as long as it doesn't break anything else or leave the code too crufty with ifdefs.
Description
Intel C++ compiler (icpc) should be supported, together with the "-xHost" compiler option to generate instructions for the highest instruction set and processor available on the compilation host machine.
Example
In
make/default_compiler_options
setCurrently, since the optimization flag only sets the integer optimization level, there is no mechanism for adding other optimization options (e.g. for code generation).
Expected Output
Example with Intel Composer XE 2019u1:
Current Version:
v2.18.1