Polyhedron Benchmarks (Intel+Linux Sept 2015)

Linux/64 on Intel Processor

Absoft
15.0.2
Absoft(AP)
15.0.2
gfortran
5.2.0
Intel
16.0
Intel(AP)
16.0
NAG
6.0
Oracle
12.4
PGI
15.7
open64
4.5.2.1
AC 4.74 4.79 6.37 4.47 4.43 8.25 9.15 6.49 4.66
AERMOD 10.56 10.97 15.63 11.54 11.62 17.26 11.34 11.19 14.69
AIR 3.32 2.76 3.17 2.57 1.65 4.37 2.86 2.90 3.25
CAPACITA 19.08 17.38 19.20 17.04 16.89 20.28 21.18 14.35 17.87
CHANNEL2 71.72 30.85 82.80 83.73 28.46 104.58 82.94 79.46 102.71
DODUC 18.31 18.25 18.85 15.03 15.03 21.13 16.04 17.00 17.52
FATIGUE2 60.79 61.72 67.12 55.83 55.35 80.76 89.40 85.73 73.42
GAS_DYN2 63.35 47.71 84.46 69.12 36.78 174.01 92.61 109.25 64.74
INDUCT2 79.33 58.54 80.91 72.03 54.06 125.36 138.32 120.72 136.02
LINPK 5.00 5.17 4.86 5.18 4.39 7.16 4.67 5.76 5.56
MDBX 9.17 8.79 8.07 6.49 4.95 8.69 8.23 8.57 8.80
MP_PROP_DESIGN 109.44 13.84 157.87 62.80 11.88 223.56 179.47 83.38 119.99
NF 7.77 8.28 7.28 7.47 7.57 8.51 9.06 8.07 7.59
PROTEIN 20.48 20.96 21.12 21.20 21.59 21.84 21.64 21.43 20.26
RNFLOW 14.70 20.54 13.68 12.90 10.13 15.39 17.46 16.05 20.38
TEST_FPU2 57.89 40.38 50.22 43.67 38.41 72.91 65.48 45.82 53.88
TFFT2 57.62 56.24 47.44 58.51 63.20 57.37 56.61 54.72 57.09
Geometric Mean 21.63 17.41 22.94 19.52 14.70 29.20 25.54 22.91 23.85
Compiler Switches
Absoft af90 -m64 -mavx -O4 -speed_math=10 -march=core -xINTEGER
Absoft(AP) af90 -m64 -mavx -O5 -speed_math=10 -march=core -xINTEGER  (These settings enable auto-parallelization, see below)
gfortran gfortran -ffast-math -funroll-loops –param max-unroll-times=2 -Ofast -march=native
Intel ifort -O3 -fast -ipo
Intel(AP) ifort -O3 -fast -parallel -ipo (These settings enable auto-parallelization, see below)
NAG nagfor -O4 -s -v -V
PGI pgf95 -V -fast -Mipa=fast,inline -Msmartalloc -Mfprelaxed -Mstack_arrays
Oracle
sunf95 -fast -xtarget=native -m64
Open64
openf95 -Ofast
Notes
All figures are execution times in seconds – measured on a machine with 2 x Xeon E5-2643 3.30GHz quad core processors, running at stock speed with hyperthreading disabled, 64 GBytes memory, and running 64-bit Linux Mint 17.1. Each figure is the average over at least 10 runs (many more for some). Measurement error is typically <1%. Green cells highlight figures within 10% of the fastest. Red cells indicate figures which are more than 150% of the fastest. So far as possible, we have used the compiler switches which give the best overall results. We have not attempted to tune individual benchmarks, and, in particular cases, different switch settings may give better results. The settings used for the Intel and Absoft compilers enable autoparallelization. Autoparallelization settings are not used on any other compilers because we found that they produced no significant performance benefits on this benchmark set. Thanks are due to Jos Bergervoet for permission to use his CAPACITA benchmark, to Quetzal Associates for permission to use their CHANNEL, FATIGUE, GAS_DYN, INDUCT, PROTEIN and RNFLOW benchmarks, to David Frank for his TEST_FPU benchmark, to Anthony Falzone for the use of MP_PROP_DESIGN, and to Ted Addison of McVehil-Monnett Associates for permission to use AERMOD, an air quality model used by the US Environmental Protection Agency. All the benchmarks have been modified slightly to fit into our benchmarking harness.