SPEC® ACCEL™ OMP Result

Copyright 2015-2017 Standard Performance Evaluation Corporation

Intel

Intel Xeon Gold 6148

Intel Server System R2208WFTZS (2 x Intel Xeon
Gold
6148, 2.40 GHz, SMT ON, Turbo ON) Endeavour Node

SPECaccel_omp_base = 5.16

SPECaccel_omp_peak = 5.61

ACCEL license: 13 Test date: Jul-2017
Test sponsor: Intel Hardware Availability: Jul-2017
Tested by: Intel Software Availability: Oct-2017
Benchmark results graph
Hardware
CPU Name: Intel Xeon Gold 6148
CPU Characteristics: Simultaneous multithreading (SMT) ON, Turbo ON
CPU MHz: 2400
CPU MHz Maximum: 3700
FPU: Integrated
CPU(s) enabled: 40 cores, 2 chips, 20 cores/chip, 2 threads/core
CPU(s) orderable: 1,2 chips
Primary Cache: 32 KB I + 32 KB D on chip per core
Secondary Cache: 1 MB I+D on chip per core
L3 Cache: 27.5 MB I+D on chip per chip
Other Cache: None
Memory: 196 GB (12 x 16 GB 2Rx4 DDR4-2666
ECC Registered)
Disk Subsystem: 108 TB Panasas ActiveStor 14
Other Hardware: None
Accelerator
Accel Model Name: Intel Xeon Gold 6148
Accel Vendor: Intel
Accel Name: Intel Xeon Gold 6148
Type of Accel: CPU
Accel Connection: N/A
Does Accel Use ECC: yes
Accel Description: 2 x Intel Xeon Gold 6148, 2.40 GHz, SMT ON, Turbo
ON
Accel Driver: N/A
Software
Operating System: Red Hat Enterprise Linux Server release 7.3
(Maipo)
3.10.0-514.6.2.0.1.el7.x86_64.knl1
Compiler: C/C++/Fortran: Version 18.0 of Intel Composer XE
for Linux Build
File System: panfs
System State: Run level 3 (default)
Other Software: FFTW 3.3.6

Results Table

Benchmark Base Peak
Seconds Ratio Seconds Ratio Seconds Ratio Seconds Ratio Seconds Ratio Seconds Ratio
Results appear in the order in which they were run. Bold underlined text indicates a median measurement.
503.postencil 43.7  2.49  45.7  2.39  44.7  2.44  43.7  2.49  45.7  2.39  44.7  2.44 
504.polbm 39.1  3.12  41.1  2.97  40.3  3.03  39.1  3.12  41.1  2.97  40.3  3.03 
514.pomriq 180    3.46  188    3.30  182    3.42  180    3.46  188    3.30  182    3.42 
550.pmd 51.0  4.73  51.0  4.73  51.7  4.66  46.9  5.14  46.7  5.16  46.6  5.17 
551.ppalm 210    2.59  209    2.60  210    2.59  102    5.33  99.2  5.48  103    5.27 
552.pep 72.2  3.20  69.7  3.32  70.0  3.30  65.1  3.55  65.2  3.54  65.0  3.55 
553.pclvrleaf 166    6.89  167    6.85  166    6.89  143    8.00  142    8.06  143    8.01 
554.pcg 52.9  6.29  54.1  6.16  53.0  6.28  52.9  6.29  54.1  6.16  53.0  6.28 
555.pseismic 92.5  3.05  90.9  3.10  91.1  3.10  87.5  3.22  87.7  3.21  85.7  3.29 
556.psp 52.7  15.5   56.0  14.6   52.7  15.5   52.7  15.5   56.0  14.6   52.7  15.5  
557.pcsp 53.9  15.9   55.3  15.5   53.6  16.0   53.9  15.9   55.3  15.5   53.6  16.0  
559.pmniGhost 93.4  4.25  94.3  4.21  93.1  4.27  83.1  4.78  83.1  4.78  83.2  4.77 
560.pilbdc 155    4.22  154    4.24  155    4.21  152    4.30  153    4.28  152    4.31 
563.pswim 47.3  3.36  49.0  3.25  47.0  3.38  45.2  3.52  45.3  3.51  45.2  3.52 
570.pbt 32.0  24.4   32.2  24.2   32.9  23.7   32.0  24.4   32.2  24.2   32.9  23.7  

Platform Notes

 Sysinfo program
 /panfs/projects/innl/abobyr/SpecACCEL_OMP/kits/kit1.2_skx_18.0/Docs/sysinfo
 $Rev: 6965 $ $Date:: 2015-04-21 #$ c05a7f14b1b1765e3fe1df68447e8a35
 running on epb227 Tue Jul 18 06:09:41 2017

 This section contains SUT (System Under Test) info as seen by
 some common utilities.  To remove or add to this section, see:
   http://www.spec.org/accel/Docs/config.html#sysinfo

 From /proc/cpuinfo
    model name : Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
       2 "physical id"s (chips)
       80 "processors"
    cores, siblings (Caution: counting these is hw and system dependent.  The
    following excerpts from /proc/cpuinfo might not be reliable.  Use with
    caution.)
       cpu cores : 20
       siblings  : 40
       physical 0: cores 0 1 2 3 4 8 9 10 11 12 16 17 18 19 20 24 25 26 27 28
       physical 1: cores 0 1 2 3 4 8 9 10 11 12 16 17 18 19 20 24 25 26 27 28
    cache size : 28160 KB

 From /proc/meminfo
    MemTotal:       196699188 kB
    HugePages_Total:       0
    Hugepagesize:       2048 kB

 From /etc/*release* /etc/*version*
    oracle-release: Oracle Linux Server release 7.3
    os-release:
       NAME="Oracle Linux Server"
       VERSION="7.3"
       ID="ol"
       VERSION_ID="7.3"
       PRETTY_NAME="Oracle Linux Server 7.3"
       ANSI_COLOR="0;31"
       CPE_NAME="cpe:/o:oracle:linux:7:3:server"
       HOME_URL="https://linux.oracle.com/"
    redhat-release: Red Hat Enterprise Linux Server release 7.3 (Maipo)
    system-release: Oracle Linux Server release 7.3
    system-release-cpe: cpe:/o:oracle:linux:7:3:server

 uname -a:
    Linux epb227 3.10.0-514.6.2.0.1.el7.x86_64.knl1 #1 SMP Thu Mar 2 10:19:17 MST
    2017 x86_64 x86_64 x86_64 GNU/Linux

 run-level 3 Jul 17 12:09

 SPEC is set to: /panfs/projects/innl/abobyr/SpecACCEL_OMP/kits/kit1.2_skx_18.0
    Filesystem                Type   Size  Used Avail Use% Mounted on
    panfs://36.101.212.1/innl panfs  108T   41T   68T  38% /global/panfs02/innl
 Additional information from dmidecode:

    Warning: Use caution when you interpret this section. The 'dmidecode' program
    reads system data which is "intended to allow hardware to be accurately
    determined", but the intent may not be met, as there are frequent changes to
    hardware, firmware, and the "DMTF SMBIOS" standard.


 (End of data from sysinfo program)

General Notes

Used Environment Variables:
  ENV_KMP_AFFINITY=compact,0 - assign OpenMP Threads continously
  ENV_OMP_NUM_THREADS=80 - limits number of Threads to be started
  ENV_KMP_HW_SUBSET=2S,20C,2T - control Thread distribution accross sockets, cores and hw threads
  ENV_FORT_BUFFERED=true - enables buffered I/O for Fortran
  ENV_OMP_DYNAMIC=FALSE - disable the dynamic adjustment of the number of threads within a team
  ENV_KMP_LIBRARY=turnaround - selects the OpenMP runtime library throughput
  ENV_KMP_BLOCKTIME=infinite - sets the time, in milliseconds, that a thread should wait,
  after completing the execution of a parallel region, before sleeping.

Base Compiler Invocation

C benchmarks:

 icc 

Fortran benchmarks:

 ifort 

Benchmarks using both Fortran and C:

 icc   ifort 

Base Portability Flags

503.postencil:  -DSPEC_USE_INNER_SIMD 
504.polbm:  -DSPEC_USE_INNER_SIMD 
514.pomriq:  -DSPEC_USE_INNER_SIMD 
550.pmd:  -DSPEC_USE_INNER_SIMD   -80 
551.ppalm:  -DSPEC_USE_INNER_SIMD 
552.pep:  -DSPEC_USE_INNER_SIMD 
553.pclvrleaf:  -DSPEC_USE_INNER_SIMD 
554.pcg:  -DSPEC_USE_INNER_SIMD 
555.pseismic:  -DSPEC_USE_INNER_SIMD 
556.psp:  -DSPEC_USE_INNER_SIMD 
557.pcsp:  -DSPEC_USE_INNER_SIMD 
559.pmniGhost:  -DSPEC_USE_INNER_SIMD   -nofor-main 
560.pilbdc:  -DSPEC_USE_INNER_SIMD 
563.pswim:  -DSPEC_USE_INNER_SIMD 
570.pbt:  -DSPEC_USE_INNER_SIMD 

Base Optimization Flags

C benchmarks:

 -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host 

Fortran benchmarks:

 -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host 

Benchmarks using both Fortran and C:

 -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host 

Peak Compiler Invocation

C benchmarks:

 icc 

Fortran benchmarks:

 ifort 

Benchmarks using both Fortran and C:

 icc   ifort 

Peak Portability Flags

503.postencil:  -DSPEC_USE_INNER_SIMD 
504.polbm:  -DSPEC_USE_INNER_SIMD 
514.pomriq:  -DSPEC_USE_INNER_SIMD 
550.pmd:  -DSPEC_USE_INNER_SIMD   -80 
551.ppalm:  -DSPEC_USE_INNER_SIMD   -DSPEC_HOST_FFTW3 
552.pep:  -DSPEC_USE_INNER_SIMD 
553.pclvrleaf:  -DSPEC_USE_INNER_SIMD 
554.pcg:  -DSPEC_USE_INNER_SIMD 
555.pseismic:  -DSPEC_USE_INNER_SIMD 
556.psp:  -DSPEC_USE_INNER_SIMD 
557.pcsp:  -DSPEC_USE_INNER_SIMD 
559.pmniGhost:  -DSPEC_USE_INNER_SIMD   -nofor-main 
560.pilbdc:  -DSPEC_USE_INNER_SIMD 
563.pswim:  -DSPEC_USE_INNER_SIMD 
570.pbt:  -DSPEC_USE_INNER_SIMD 

Peak Optimization Flags

C benchmarks:

503.postencil:  basepeak = yes 
504.polbm:  basepeak = yes 
514.pomriq:  basepeak = yes 
552.pep:  -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host   -qopt-streaming-stores always   -fimf-precision=low 
554.pcg:  basepeak = yes 
557.pcsp:  basepeak = yes 
570.pbt:  basepeak = yes 

Fortran benchmarks:

550.pmd:  -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host   -fimf-precision=low 
551.ppalm:  -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host   -I/home/abobyr/FFTW-3.3.6_SKL/include   -L/home/abobyr/FFTW-3.3.6_SKL/lib 
555.pseismic:  -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host 
556.psp:  basepeak = yes 
560.pilbdc:  -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host   -qopt-prefetch=5 
563.pswim:  Same as 555.pseismic 

Benchmarks using both Fortran and C:

 -O3   -xCOMMON-AVX512   -qopenmp   -qopenmp-offload=host   -qopt-streaming-stores always 

Peak Other Flags

Fortran benchmarks:

551.ppalm:  -lfftw3 

The flags file that was used to format this result can be browsed at
https://www.spec.org/accel/flags/Intel-icc18.0-linux64.20170802.html.

You can also download the XML flags source by saving the following link:
https://www.spec.org/accel/flags/Intel-icc18.0-linux64.20170802.xml.