SPEC® CINT2006 Result

Copyright 2006-2017 Standard Performance Evaluation Corporation

Sugon

Sugon A620-G30 (AMD EPYC 7451)

CPU2006 license: 9046 Test date: Dec-2017
Test sponsor: Sugon Hardware Availability: Dec-2017
Tested by: Sugon Software Availability: Oct-2017
Benchmark results graph
Hardware
CPU Name: AMD EPYC 7451
CPU Characteristics: AMD Turbo CORE technology up to 3.20 GHz
CPU MHz: 2300
FPU: Integrated
CPU(s) enabled: 48 cores, 2 chips, 24 cores/chip, 2 threads/core
CPU(s) orderable: 1,2 chips
Primary Cache: 64 KB I + 32 KB D on chip per core
Secondary Cache: 512 KB I+D on chip per core
L3 Cache: 64 MB I+D on chip per chip, 8 MB shared / 3 cores
Other Cache: None
Memory: 1 TB (16 x 64 GB 2S2Rx4 PC4-2667V-R, running
at 2400)
Disk Subsystem: 1 x 2000 GB SATA, 7200 RPM
Other Hardware: None
Software
Operating System: Red Hat Enterprise Linux Server 7.4
Kernel 3.10.0-693.2.2
Compiler: C/C++: Version 4.5.2.1 of x86 Open64 Compiler
Suite (from AMD)
Auto Parallel: No
File System: ext4
System State: Run level 3 (Multi User)
Base Pointers: 32/64-bit
Peak Pointers: 32/64-bit
Other Software: MicroQuill SmartHeap 10.0 32-bit Library for Linux

Results Table

Benchmark Base Peak
Copies Seconds Ratio Seconds Ratio Seconds Ratio Copies Seconds Ratio Seconds Ratio Seconds Ratio
Results appear in the order in which they were run. Bold underlined text indicates a median measurement.
400.perlbench 96 591 1590 594 1580 591 1590 96 469 2000 469 2000 468 2000
401.bzip2 96 864 1070 869 1070 868 1070 96 836 1110 838 1110 836 1110
403.gcc 96 596 1300 598 1290 598 1290 96 539 1430 538 1440 538 1440
429.mcf 96 596 1470 593 1480 595 1470 96 386 2270 383 2290 386 2270
445.gobmk 96 688 1460 688 1460 687 1470 96 647 1560 647 1560 646 1560
456.hmmer 96 302 2970 303 2950 303 2960 96 295 3030 282 3180 281 3190
458.sjeng 96 854 1360 853 1360 852 1360 96 818 1420 816 1420 818 1420
462.libquantum 96 110 18200 110 18100 110 18100 96 112 17800 112 17800 112 17800
464.h264ref 96 1055 2010 1054 2020 1061 2000 96 877 2420 875 2430 875 2430
471.omnetpp 96 615 976 615 976 616 974 96 630 952 626 958 624 961
473.astar 96 609 1110 611 1100 612 1100 96 551 1220 550 1230 549 1230
483.xalancbmk 96 333 1990 333 1990 334 1980 96 333 1990 331 2000 332 2000

Submit Notes

The config file option 'submit' was used.
'numactl' was used to bind copies to the cores.
See the configuration file for details.

Operating System Notes

'ulimit -s unlimited' was used to set environment stack size
'ulimit -l 2097152' was used to set environment locked pages in memory limit

runspec command invoked through numactl i.e.:
numactl --interleave=all runspec <etc>

Set dirty_ratio=8 to limit dirty cache to 8% of memory
Set swappiness=1 to swap only if necessary
Set zone_reclaim_mode=1 to free local node memory and avoid remote memory
sync then drop_caches=3 to reset caches before invoking runcpu

Transparent huge pages were enabled for this run (OS default)

Set vm/nr_hugepages=86016 in /etc/sysctl.conf
mount -t hugetlbfs nodev /mnt/hugepages

General Notes

Environment variables set by runspec before the start of the run:
HUGETLB_LIMIT = "896"
LD_LIBRARY_PATH = "/home/cpu2006/amd1603-rate-libs-revB/32:/home/cpu2006/amd1603-rate-libs-revB/64"

The binaries were built with the AMD supported x86 Open64 Compiler Suite,
which is only available from AMD at
http://developer.amd.com/tools-and-sdks/cpu-development/x86-open64-compiler-suite/
Binaries were compiled on a system with 2 x AMD Opteron 6378 chips + 128 GB Memory using RHEL 6.3

Base Compiler Invocation

C benchmarks:

 opencc 

C++ benchmarks:

 openCC 

Base Portability Flags

400.perlbench:  -DSPEC_CPU_LP64   -DSPEC_CPU_LINUX_X64 
401.bzip2:  -DSPEC_CPU_LP64 
403.gcc:  -DSPEC_CPU_LP64 
429.mcf:  -DSPEC_CPU_LP64 
445.gobmk:  -DSPEC_CPU_LP64 
456.hmmer:  -DSPEC_CPU_LP64 
458.sjeng:  -DSPEC_CPU_LP64 
462.libquantum:  -DSPEC_CPU_LP64   -DSPEC_CPU_LINUX 
464.h264ref:  -DSPEC_CPU_LP64 
483.xalancbmk:  -DSPEC_CPU_LINUX 

Base Optimization Flags

C benchmarks:

 -Ofast   -CG:local_sched_alg=1   -INLINE:aggressive=ON   -IPA:plimit=8000   -IPA:small_pu=100   -HP:bd=2m:heap=2m   -mso   -LNO:prefetch=2   -march=bdver1   -mno-fma4   -mno-xop   -mno-tbm 

C++ benchmarks:

 -Ofast   -m32   -INLINE:aggressive=on   -CG:cmp_peep=on   -D__OPEN64_FAST_SET   -march=bdver1   -mno-fma4   -mno-xop   -mno-tbm   -L/root/work/libraries/SmartHeap-10/lib -lsmartheap 

Peak Compiler Invocation

C benchmarks:

 opencc 

C++ benchmarks:

 openCC 

Peak Portability Flags

400.perlbench:  -DSPEC_CPU_LP64   -DSPEC_CPU_LINUX_X64 
401.bzip2:  -DSPEC_CPU_LP64 
445.gobmk:  -DSPEC_CPU_LP64 
456.hmmer:  -DSPEC_CPU_LP64 
458.sjeng:  -DSPEC_CPU_LP64 
462.libquantum:  -DSPEC_CPU_LP64   -DSPEC_CPU_LINUX 
464.h264ref:  -DSPEC_CPU_LP64 
473.astar:  -DSPEC_CPU_LP64 
483.xalancbmk:  -DSPEC_CPU_LINUX 

Peak Optimization Flags

C benchmarks:

400.perlbench:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -LNO:prefetch=2   -LNO:opt=0   -IPA:plimit=20000   -OPT:unroll_times_max=8   -OPT:unroll_size=256   -OPT:unroll_level=2   -OPT:keep_ext=on   -WOPT:if_conv=0   -WOPT:sib=on   -CG:local_sched_alg=1   -CG:unroll_fb_req=on   -CG:movext_icmp=off   -HP:bd=2m:heap=2m   -march=bdver1   -mno-fma4   -GRA:aggr_loop_splitting=off   -GRA:loop_splitting=off 
401.bzip2:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -LNO:prefetch=2   -LNO:pf2=0   -OPT:alias=disjoint   -OPT:goto=off   -CG:local_sched_alg=1   -HP:bdt=2m:heap=2m   -march=bdver2   -WB,   -mno-fma4   -mno-tbm   -mno-xop 
403.gcc:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -LNO:trip_count=256   -CG:cmp_peep=on   -CG:pre_minreg_level=2   -m32   -HP:bdt=2m:heap=2m   -GRA:unspill=on   -IPA:small_pu=200   -WOPT:sib=on   -march=bdver2   -mno-fma4   -WB,   -mno-tbm   -mno-xop 
429.mcf:  -O3   -OPT:unroll_times_max=5   -ipa   -INLINE:aggressive=on   -CG:gcm=off   -CG:dsched=on   -GRA:prioritize_by_density=on   -m32   -HP:bdt=2m:heap=2m   -mso   -march=bdver1   -mno-fma4 
445.gobmk:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -OPT:unroll_size=256   -OPT:unroll_times_max=8   -OPT:keep_ext=on   -IPA:plimit=750   -IPA:min_hotness=300   -IPA:pu_reorder=1   -LNO:ignore_feedback=off   -WOPT:if_conv=2   -HP:bd=2m:heap=2m   -march=bdver1   -mno-fma4 
456.hmmer:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -LNO:prefetch=2   -OPT:alias=disjoint   -OPT:unroll_times_max=16   -OPT:unroll_size=512   -OPT:unroll_level=2   -OPT:keep_ext=on   -CG:cflow=0   -CG:cmp_peep=on   -CG:pre_local_sched=off   -HP:bdt=2m:heap=2m   -CG:p2align=0   -CG:load_exe=3   -CG:dsched=on   -march=bdver1   -mno-fma4 
458.sjeng:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -CG:ptr_load_use=0   -CG:divrem_opt=on   -CG:movext_icmp=off   -CG:locs_best=on   -LNO:full_unroll=10   -IPA:pu_reorder=2   -HP:heap=2m:bd=2m   -WOPT:sib=on   -march=bdver1   -mno-fma4 
462.libquantum:  -Ofast   -mso   -OPT:unroll_size=512   -OPT:unroll_times_max=16   -LNO:prefetch=2   -LNO:prefetch_ahead=4   -LNO:pf2=0   -CG:local_sched_alg=1   -CG:p2align=0   -INLINE:aggressive=ON   -IPA:plimit=15000   -IPA:small_pu=100   -HP:bdt=2m:heap=2m,limit=300   -march=bdver2   -mno-fma4 
464.h264ref:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -O3   -OPT:unroll_size=256   -OPT:unroll_times_max=2   -IPA:plimit=20000   -OPT:alias=disjoint   -CG:ptr_load_use=0   -CG:local_sched_alg=1   -HP:bdt=2m:heap=2m   -march=bdver1   -mno-fma4 

C++ benchmarks:

471.omnetpp:  -Ofast   -m32   -INLINE:aggressive=on   -CG:cmp_peep=on   -WOPT:sib=on   -D__OPEN64_FAST_SET   -march=bdver2   -mno-fma4   -L/root/work/libraries/SmartHeap-10/lib -lsmartheap 
473.astar:  -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -WOPT:if_conv=0   -WOPT:sib=on   -CG:divrem_opt=on   -CG:p2align=1   -CG:dsched=on   -GRA:optimize_boundary=on   -OPT:alias=disjoint   -INLINE:aggressive=on   -IPA:small_pu=3000   -IPA:plimit=3000   -HP:bdt=2m:heap=2m   -march=bdver1   -mno-fma4 
483.xalancbmk:  -Ofast   -LNO:prefetch=2   -OPT:unroll_size=512   -OPT:unroll_times_max=8   -D__OPEN64_FAST_SET   -INLINE:aggressive=on   -m32   -CG:cmp_peep=on   -CG:local_sched=off   -CG:p2align=1   -GRA:unspill=on   -TENV:frame_pointer=off   -fno-emit-exceptions   -march=bdver2   -mno-fma4   -L/root/work/libraries/SmartHeap-10/lib -lsmartheap 

The flags files that were used to format this result can be browsed at
http://www.spec.org/cpu2006/flags/x86-openflags-rate-revA-I.html,
http://www.spec.org/cpu2006/flags/Sugon-Naples-Platform-Settings-revC-I.html.

You can also download the XML flags sources by saving the following links:
http://www.spec.org/cpu2006/flags/x86-openflags-rate-revA-I.xml,
http://www.spec.org/cpu2006/flags/Sugon-Naples-Platform-Settings-revC-I.xml.