Compiling applications in Puhti
- Whenever possible, use the local disk on the login node for compiling software.
- Compiling on the local disk is much faster and shifts load from the shared file system.
- The local disk is cleaned frequently, so please move your files elsewhere after compiling.
Building CPU applications
C/C++ and Fortran applications can be built with Intel or GNU compiler suites. The compiler suite is selected via the Modules system, i.e.
module load intel
module load gcc
Different applications function better with different suites, so the selection needs to be done on a case-by-case basis.
The actual compiler commands for building a serial application with the two suites:
Intel and GNU compilers use different compiler options. The recommended basic optimization flags are listed in the table below. It is recommended to start from the safe level and then move up to intermediate or even aggressive, while making sure the results are correct and the program's performance has improved.
|Safe||-O2 -xHost -fp-model precise||-O2 -march=native|
|Intermediate||-O2 -xHost||-O3 -march=native|
|Aggressive||-O3 -xHost -fp-model fast=2 -no-prec-div -fimf-use-svml=true -qopt-zmm-usage=high||-O3 -march=native -ffast-math -funroll-loops -mprefer-vector-width=512|
Please note that not all applications benefit from the AVX-512 vector set
-march=native). It may be a good idea to also test AVX2
-mavx2) and compare the performance.
A detailed list of options for the Intel and GNU compilers can be found on the man
man gcc/gfortran when the corresponding programming
environment is loaded, or in the compiler manuals (see the links above).
List all available versions of the compiler suites:
module spider intel module spider gcc
Building GPU applications
Both the CUDA and OpenACC programming models are supported on Puhti. Specific modules have to be loaded in order to use them.
For example, to load the CUDA 10.1 environment:
module load gcc/8.3.0 cuda/10.1.168
Or to load the PGI compiler for OpenACC:
module load pgi
For more detailed information about the available modules, please see
spider cuda or
module spider pgi.
The CUDA compiler (
nvcc) takes care of compiling the CUDA code for the target
GPU device and passing on the rest to a non-CUDA compiler (
To generate code for a given target device, tell the CUDA
compiler what compute capability the target device supports. On Puhti, the
GPUs (Volta V100) support compute capability 7.0. Specify this using
For example, compiling a CUDA kernel (
example.cu) on Puhti:
nvcc -gencode arch=compute_70,code=sm_70 example.cu
In principle, it is also possible to target multiple GPU architectures by repeating
-gencode multiple times for different compute capabilities. However, this is
not necessary on Puhti, since there is only one type of GPU.
OpenACC is supported with the PGI compilers (
To enable OpenACC support, one needs to give
-acc flag to the compiler.
To generate code for a given target device, tell the compiler
what compute capability the target device supports. On Puhti, the GPUs (Volta
V100) support compute capability 7.0. Specify it with
For example, to compiling C code that uses OpenACC directives (
pgcc -acc -ta=tesla:cc70 example.c
For information about what the compiler actually does with the OpenACC
Building MPI applications
There are currently three MPI environments available: hpcx-mpi, mpich, and intel-mpi. The default is hpcx-mpi, which is also recommended to begin with.
If hpcx-mpi is incompatible with your application or delivers insufficient performance,
please try another environment. All MPI
implementations can be used with both Intel and GNU compiler suites. The PGI
compiler cannot presently be used with MPI. The MPI environments can be used
module load, i.e.
module load hpcx-mpi
When building MPI applications, use mpixxx compiler wrappers that differ depending on the compiler suite and the MPI environment:
|Compiler suite||hpcx-mpi or mpich||intel-mpi|
|Intel||mpifort, mpicc, mpicxx||mpiifort, mpiicc, mpiicpc|
|GNU||mpif90, mpicc, mpicxx||mpif90, mpicc, mpicxx|
Building OpenMP and hybrid applications
Additional compiler and linker flags are needed when building OpenMP or MPI/OpenMP hybrid applications:
|Compiler suite||OpenMP flag|