Intel Parallel Studio (Cluster Edition)
Intel Parallel Studio Cluster Edition XE 2017.2
Usage
module load intel
Getting started
Optimization
The Intel@ Advisor for Linux can help you find the required parameters for linking and compiling
file:///opt/apps/intel/2017.2/documentation_2017/en/advisor/welcomepage/get_started.htm
To find the paramters for the MKL you can use the MKL Link Line Advisor
Some Compiler Options
To get the used CPU type for which you want to optimize and some of the supported features of your CPU you can take a look at.
cat /proc/cpuinfo
New flags since Intel Compiler Version 11.1 :
| CPU Architecture Flag | -mXXXXXXX | |
| Default on Linux (and Windows) | -mSSE2 | default (icc > v12.x) |
| SSE3, SSE2 and SSE instructionsets | -mSSE3 | |
| SSSE3, Intel® SSE3, SSE2 and SSE | -mSSSE3 | |
| SSE4.1, SSSE3, SSE3, SSE2 and SSE | -mSSE4.1 | |
| SSE4.2, SSE4.1, SSSE3, SSE3, SSE2 and SSE | -mSSE4.2 | Westmere Xeon ww8-2026-1, Lima-Cluster |
| AVX, Intel® SSE4.2, SSE4.1, SSSE3, SSE3, SSE2 and SSE | -mAVX | Sandy Bridge, Ivy Bridge |
| Autoparallization | -parallel |
| use OpenMP | -openmp |
| ??? | -fPIC |
| 64bit pointer (else autodetect) | -m64 |
Optimization Flags available on OLD Intel Compilers (10.x) Listed are the only available options working with newest supported CPUs:
| Core™2 Duo processor, Xeon® 5100 series | -xT -axT |
