SPEC® CFP2006 Result

Sugon

Sugon A620-G30 (AMD EPYC 7601)

SPECfp®_rate2006 = 1850

SPECfp_rate_base2006 = 1670

CPU2006 license:	9046	Test date:	Aug-2017
Test sponsor:	Sugon	Hardware Availability:	Jul-2017
Tested by:	Sugon	Software Availability:	Apr-2017

Hardware
CPU Name:	AMD EPYC 7601
CPU Characteristics:	AMD Turbo CORE technology up to 3.20 GHz
CPU MHz:	2200
FPU:	Integrated
CPU(s) enabled:	64 cores, 2 chips, 32 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 / 4 cores
Other Cache:	None
Memory:	512 GB (16 x 32 GB 2Rx4 PC4-2666V-R, running at 2400 MHz)
Disk Subsystem:	1 x 1.0 TB SATA, 7200 RPM
Other Hardware:	None

Software
Operating System:	Ubuntu 17.04 Kernel 4.10.0-24-generic
Compiler:	C/C++/Fortran: Version 4.5.2.1 of x86 Open64 Compiler Suite (from AMD)
Auto Parallel:	No
File System:	ext4
System State:	Run level 3 (Full multiuser with network)
Base Pointers:	64-bit
Peak Pointers:	32/64-bit
Other Software:	None

Results Table

Benchmark	Base							Peak
Benchmark	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.
410.bwaves	128	1456	1190	1457	1190	1457	1190	128	1426	1220	1424	1220	1424	1220
416.gamess	128	1147	2190	1138	2200	1140	2200	128	1014	2470	1013	2470	1012	2480
433.milc	128	1204	976	1204	976	1204	976	128	1009	1160	1009	1160	1009	1160
434.zeusmp	128	538	2170	537	2170	538	2170	128	530	2200	530	2200	530	2200
435.gromacs	128	420	2170	420	2180	424	2150	128	313	2920	322	2840	311	2930
436.cactusADM	128	658	2320	656	2330	652	2350	128	644	2380	641	2390	643	2380
437.leslie3d	128	1397	861	1392	865	1392	864	128	1091	1100	1090	1100	1089	1100
444.namd	128	523	1960	523	1960	524	1960	128	449	2290	446	2300	447	2300
447.dealII	128	424	3460	423	3460	416	3520	128	371	3950	365	4010	382	3830
450.soplex	128	1150	928	1151	928	1150	928	128	1071	997	1069	998	1070	998
453.povray	128	251	2710	251	2720	248	2740	128	203	3350	200	3400	200	3400
454.calculix	128	356	2970	377	2800	363	2910	128	378	2800	378	2790	378	2790
459.GemsFDTD	128	1756	774	1755	774	1757	773	128	1631	833	1629	834	1632	832
465.tonto	128	640	1970	640	1970	643	1960	128	589	2140	588	2140	590	2130
470.lbm	128	998	1760	999	1760	997	1760	128	998	1760	998	1760	1013	1740
481.wrf	128	951	1500	952	1500	950	1500	128	954	1500	950	1510	952	1500
482.sphinx3	128	1965	1270	1965	1270	1968	1270	128	1770	1410	1766	1410	1769	1410

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 runspec

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

vm.nr_hugepages=114688 in /etc/sysctl.conf
mount -t hugetlbfs nodev /mnt/hugepages

Platform Notes

The Linux run level was 3; sysinfo run-level is incorrect.
The dmidecode infomation displayed in sysinfo should have one line reading as:
16x Samsung M393A4K40BB2-CTD 32 GB 2 rank 2666 MHz, configured at 2400 MHz
Set "SMT Mode" in the BIOS as "Auto".

General Notes

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

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

Base Compiler Invocation

C benchmarks:

opencc

C++ benchmarks:

openCC

Fortran benchmarks:

openf95

Benchmarks using both Fortran and C:

opencc openf95

Base Portability Flags

410.bwaves:	-DSPEC_CPU_LP64
416.gamess:	-DSPEC_CPU_LP64
433.milc:	-DSPEC_CPU_LP64
434.zeusmp:	-DSPEC_CPU_LP64
435.gromacs:	-DSPEC_CPU_LP64
436.cactusADM:	-DSPEC_CPU_LP64 -fno-second-underscore
437.leslie3d:	-DSPEC_CPU_LP64
444.namd:	-DSPEC_CPU_LP64
447.dealII:	-DSPEC_CPU_LP64
450.soplex:	-DSPEC_CPU_LP64
453.povray:	-DSPEC_CPU_LP64
454.calculix:	-DSPEC_CPU_LP64
459.GemsFDTD:	-DSPEC_CPU_LP64
465.tonto:	-DSPEC_CPU_LP64
470.lbm:	-DSPEC_CPU_LP64
481.wrf:	-DSPEC_CPU_LINUX -DSPEC_CPU_CASE_FLAG -DSPEC_CPU_LP64 -fno-second-underscore
482.sphinx3:	-DSPEC_CPU_LP64

Base Optimization Flags

Peak Compiler Invocation

C benchmarks:

opencc

C++ benchmarks:

openCC

Fortran benchmarks:

openf95

Benchmarks using both Fortran and C:

opencc openf95

Peak Portability Flags

410.bwaves:	-DSPEC_CPU_LP64
416.gamess:	-DSPEC_CPU_LP64
433.milc:	-DSPEC_CPU_LP64
434.zeusmp:	-DSPEC_CPU_LP64
435.gromacs:	-DSPEC_CPU_LP64
436.cactusADM:	-DSPEC_CPU_LP64 -fno-second-underscore
437.leslie3d:	-DSPEC_CPU_LP64
444.namd:	-DSPEC_CPU_LP64
453.povray:	-DSPEC_CPU_LP64
454.calculix:	-DSPEC_CPU_LP64
459.GemsFDTD:	-DSPEC_CPU_LP64
465.tonto:	-DSPEC_CPU_LP64
470.lbm:	-DSPEC_CPU_LP64
481.wrf:	-DSPEC_CPU_LINUX -DSPEC_CPU_CASE_FLAG -DSPEC_CPU_LP64 -fno-second-underscore

Peak Optimization Flags

C benchmarks:

433.milc:	-Ofast -CG:movnti=1 -CG:locs_best=on -HP:bdt=2m:heap=2m -IPA:plimit=7000 -IPA:callee_limit=1200 -OPT:struct_array_copy=2 -OPT:alias=field_sensitive -mso -march=bdver1 -mno-fma4
470.lbm:	-Ofast -CG:cmp_peep=on -OPT:keep_ext=on -HP:bdt=2m:heap=2m -IPA:plimit=8000 -IPA:small_pu=100 -march=bdver1 -mno-fma4 -mso
482.sphinx3:	-Ofast -m32 -IPA:plimit=1000 -OPT:malloc_alg=2 -CG:cmp_peep=on -CG:p2align=0 -CG:load_exe=1 -CG:dsched=on -INLINE:aggressive=on -LNO:prefetch=2 -LNO:prefetch_ahead=4 -mso -march=bdver2 -WB, -mno-fma4 -mno-tbm -mno-xop

C++ benchmarks:

444.namd:	-Ofast -IPA:plimit=3000 -LNO:ignore_feedback=off -CG:local_sched_alg=0 -CG:load_exe=0 -OPT:unroll_size=256 -fno-exceptions -HP:bdt=2m:heap=2m -LNO:if_select_conv=1 -OPT:alias=disjoint -LNO:psimd_iso_unroll=ON -march=bdver2 -mno-fma4 -WB, -mno-xop -mno-tbm
447.dealII:	-Ofast -D__OPEN64_FAST_SET -static -INLINE:aggressive=on -LNO:opt=1 -LNO:simd=2 -fno-emit-exceptions -m32 -OPT:unroll_times_max=8 -OPT:unroll_size=256 -OPT:unroll_level=2 -HP:bdt=2m:heap=2m -GRA:unspill=on -CG:cmp_peep=on -CG:movext_icmp=off -TENV:frame_pointer=off -march=bdver1 -mno-fma4
450.soplex:	-fb_create fbdata(pass 1) -fb_opt fbdata(pass 2) -O3 -LNO:ignore_feedback=off -INLINE:aggressive=on -OPT:RO=1 -OPT:IEEE_arith=3 -OPT:IEEE_NaN_Inf=off -OPT:fold_unsigned_relops=on -fno-exceptions -CG:p2align=0 -m32 -mno-fma4 -HP:bdt=2m:heap=2m -WOPT:sib=on -march=bdver1
453.povray:	-fb_create fbdata(pass 1) -fb_opt fbdata(pass 2) -Ofast -CG:pre_local_sched=off -CG:p2align=0 -CG:p2align_split=on -CG:dsched=on -INLINE:aggressive=on -HP:bd=2m:heap=2m -OPT:transform=2 -OPT:alias=disjoint -WOPT:aggcm=0 -march=bdver2 -mno-fma4 -WB, -mno-xop -mno-tbm -Wl, -z,muldefs

Fortran benchmarks:

410.bwaves:	-fb_create fbdata(pass 1) -fb_opt fbdata(pass 2) -Ofast -OPT:Ofast -OPT:treeheight=on -LNO:blocking=off -LNO:ignore_feedback=off -LNO:fu=4 -LNO:loop_model_simd=on -LNO:simd_rm_unity_remainder=on -WOPT:aggstr=0 -HP:bdt=2m:heap=2m -CG:cmp_peep=on -march=bdver2 -mno-fma4
416.gamess:	-fb_create fbdata(pass 1) -fb_opt fbdata(pass 2) -Ofast -LNO:fu=6 -LNO:blocking=0 -LNO:simd=2 -OPT:ro=3 -OPT:recip=on -CG:local_sched_alg=1 -HP:bdt=2m:heap=2m -WOPT:sib=on -march=bdver1 -mno-fma4
434.zeusmp:	-fb_create fbdata(pass 1) -fb_opt fbdata(pass 2) -Ofast -LNO:blocking=off -LNO:interchange=off -IPA:plimit=1500 -HP:bdt=2m:heap=2m -march=bdver2 -mno-fma4
437.leslie3d:	-Ofast -CG:pre_minreg_level=2 -LNO:simd=0 -LNO:fusion=2 -HP:bdt=2m:heap=2m -mso -march=bdver1 -mno-fma4
459.GemsFDTD:	-Ofast -IPA:plimit=1500 -OPT:unroll_size=1024 -OPT:unroll_times_max=16 -LNO:fission=2 -CG:local_sched_alg=2 -HP -march=bdver1 -mno-fma4
465.tonto:	-Ofast -OPT:alias=no_f90_pointer_alias -LNO:blocking=off -CG:load_exe=1 -CG:local_sched_alg=3 -IPA:plimit=525 -HP:bdt=2m:heap=2m -march=bdver2 -WB, -mno-fma4 -mno-tbm -mno-xop

Benchmarks using both Fortran and C:

435.gromacs:	-Ofast -OPT:rsqrt=2 -HP:bdt=2m:heap=2m -CG:local_sched_alg=2 -CG:load_exe=3 -GRA:unspill=on -march=bdver2 -mno-fma4 -LNO:simd=3
436.cactusADM:	-fb_create fbdata(pass 1) -fb_opt fbdata(pass 2) -Ofast -LNO:blocking=off -LNO:prefetch=2 -LNO:pf2=0 -LNO:prefetch_ahead=4 -HP -CG:locs_shallow_depth=1 -CG:load_exe=0 -CG:dsched=on -WOPT:sib=on -march=bdver2 -mno-fma4
454.calculix:	-Ofast -OPT:unroll_size=256 -OPT:alias=disjoint -GRA:optimize_boundary=on -CG:dsched=on -HP:bdt=2m:heap=2m -march=bdver1 -mno-fma4
481.wrf:	-Ofast -LNO:blocking=off -LANG:copyinout=off -IPA:callee_limit=5000 -GRA:prioritize_by_density=on -HP -WOPT:sib=on -march=bdver1 -mno-fma4

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-revB-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-revB-I.xml.