| CPU2006 license: | 6 | Test date: | Mar-2013 |
|---|---|---|---|
| Test sponsor: | Oracle Corporation | Hardware Availability: | Mar-2013 |
| Tested by: | Oracle Corporation | Software Availability: | Feb-2013 |
| Hardware | |
|---|---|
| CPU Name: | SPARC T5 |
| CPU Characteristics: | |
| CPU MHz: | 3600 |
| FPU: | Integrated |
| CPU(s) enabled: | 128 cores, 8 chips, 16 cores/chip, 8 threads/core |
| CPU(s) orderable: | 8 chips |
| Primary Cache: | 16 KB I + 16 KB D on chip per core |
| Secondary Cache: | 128 KB I+D on chip per core |
| L3 Cache: | 8 MB I+D on chip per chip |
| Other Cache: | None |
| Memory: | 4 TB (128 x 32 GB), 4-way interleaved |
| Disk Subsystem: | 2 TB on 8x 600 GB 10K RPM SAS disks, arranged as 4x 2-way mirrors |
| Other Hardware: | None |
| Software | |
|---|---|
| Operating System: | Oracle Solaris 11.1, SRU 4.6 |
| Compiler: | C/C++/Fortran: Version 12.3 of Oracle Solaris, 1/13 Platform Specific Enhancement |
| Auto Parallel: | No |
| File System: | zfs with gzip compression |
| System State: | Default |
| Base Pointers: | 32-bit |
| Peak Pointers: | 32-bit |
| Other Software: | None |
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. | ||||||||||||||
| 410.bwaves | 640 | 2783 | 3130 | 2774 | 3140 | 2742 | 3170 | 384 | 1623 | 3220 | 1680 | 3110 | 1622 | 3220 |
| 416.gamess | 640 | 5326 | 2350 | 5337 | 2350 | 5327 | 2350 | 640 | 4803 | 2610 | 4802 | 2610 | 4800 | 2610 |
| 433.milc | 640 | 1669 | 3520 | 1688 | 3480 | 1669 | 3520 | 1016 | 2671 | 3490 | 2612 | 3570 | 2614 | 3570 |
| 434.zeusmp | 640 | 1859 | 3130 | 1869 | 3120 | 1863 | 3130 | 640 | 1859 | 3130 | 1869 | 3120 | 1863 | 3130 |
| 435.gromacs | 640 | 2600 | 1760 | 2599 | 1760 | 2597 | 1760 | 1016 | 3923 | 1850 | 3897 | 1860 | 3898 | 1860 |
| 436.cactusADM | 640 | 2346 | 3260 | 2344 | 3260 | 2344 | 3260 | 384 | 1354 | 3390 | 1355 | 3390 | 1355 | 3390 |
| 437.leslie3d | 640 | 2896 | 2080 | 2896 | 2080 | 2903 | 2070 | 512 | 2284 | 2110 | 2286 | 2110 | 2288 | 2100 |
| 444.namd | 640 | 1787 | 2870 | 1791 | 2870 | 1788 | 2870 | 896 | 2391 | 3000 | 2429 | 2960 | 2386 | 3010 |
| 447.dealII | 640 | 1691 | 4330 | 1700 | 4310 | 1689 | 4330 | 640 | 1691 | 4330 | 1700 | 4310 | 1689 | 4330 |
| 450.soplex | 640 | 3930 | 1360 | 4241 | 1260 | 4041 | 1320 | 1016 | 3698 | 2290 | 3709 | 2280 | 3705 | 2290 |
| 453.povray | 640 | 871 | 3910 | 867 | 3930 | 864 | 3940 | 640 | 795 | 4280 | 794 | 4290 | 786 | 4330 |
| 454.calculix | 640 | 2012 | 2620 | 2012 | 2620 | 2010 | 2630 | 512 | 1559 | 2710 | 1556 | 2720 | 1557 | 2710 |
| 459.GemsFDTD | 640 | 3102 | 2190 | 3104 | 2190 | 3102 | 2190 | 384 | 1768 | 2300 | 1769 | 2300 | 1768 | 2300 |
| 465.tonto | 640 | 2280 | 2760 | 2290 | 2750 | 2281 | 2760 | 640 | 2224 | 2830 | 2219 | 2840 | 2225 | 2830 |
| 470.lbm | 640 | 2255 | 3900 | 2256 | 3900 | 2260 | 3890 | 384 | 1027 | 5140 | 1029 | 5130 | 1026 | 5140 |
| 481.wrf | 640 | 2411 | 2970 | 2432 | 2940 | 2410 | 2970 | 640 | 2411 | 2970 | 2432 | 2940 | 2410 | 2970 |
| 482.sphinx3 | 640 | 4199 | 2970 | 4207 | 2970 | 4201 | 2970 | 896 | 5311 | 3290 | 5300 | 3290 | 5302 | 3290 |
Submit Notes
Processes were bound to cores using "submit" and "pbind". The config file option 'submit' was used in order to accomplish this.
Operating System Notes
ulimit -s 131072 was used to limit the space
consumed by the stack
/etc/system parameters
autoup=600
Causes pages older than the listed number of seconds to
be written by fsflush.
tune_t_fsflushr=10
Controls how many seconds elapse between runs of the
page flush daemon, fsflush.
lpg_alloc_prefer=1
Indicates that extra effort should be taken to ensure
that pages are created in the nearby lgroup (NUMA location).
tsb_rss_factor=128
Suggests that the the size of the TSB (Translation
Storage Buffer) may be increased if it is more than
25% (128/512) full. Doing so may reduce TSB traps,
at the cost of additional kernel memory.
zfs:zfs_arc_min = 1073741824
zfs:zfs_arc_max = 0x10000000000
Limits the consumption of memory by the zfs file system
cache to the range 1 GB to 1 TB.
plat_disable_mempm=1
Do not attempt to allocate kernel memory in a power-aware fashion
poweradm set administrative-authority=none
Disables Solaris power management
A swapfile of 4 TB was connected via Fibre Channel
on a Sun Storage 2540-M2 with 12x SAS 600 GB
disks, Raid-5
Platform Notes
Power policy set to 'disabled' at ILOM Power Managment menu Sysinfo program /cpu2006/config/sysinfo.rev6818 $Rev: 6818 $ $Date:: 2012-07-17 #$ 5569a0425e2ad530534e4c79a46e4d28 running on bur407-139.us.oracle.com Mon Mar 18 12:54:17 2013 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/cpu2006/Docs/config.html#sysinfo From /usr/sbin/psrinfo SPARC-T5 (chipid 0, clock 3600 MHz) SPARC-T5 (chipid 1, clock 3600 MHz) SPARC-T5 (chipid 2, clock 3600 MHz) SPARC-T5 (chipid 3, clock 3600 MHz) SPARC-T5 (chipid 4, clock 3600 MHz) SPARC-T5 (chipid 5, clock 3600 MHz) SPARC-T5 (chipid 6, clock 3600 MHz) SPARC-T5 (chipid 7, clock 3600 MHz) 8 chips 1024 threads 3600 MHz From kstat: 128 cores From prtconf: 4 Terabytes /etc/release: Oracle Solaris 11.1 SPARC uname -a: SunOS bur407-139.us.oracle.com 5.11 11.1 sun4v sparc sun4v disk: df -h $SPEC Filesystem Size Used Available Capacity Mounted on rpool/export/cpu2006 98G 973M 86G 2% /cpu2006 (End of data from sysinfo program)
Base Compiler Invocation
C benchmarks:
| cc |
C++ benchmarks:
| CC |
Fortran benchmarks:
| f90 |
Benchmarks using both Fortran and C:
| cc f90 |
(*) Indicates a compiler flag that was found in a non-compiler variable.
Base Optimization Flags
C benchmarks:
| -g -fast -fma=fused -xipo=2 -xpagesize=4M -xprefetch_level=2 -xalias_level=std -xprefetch_level=3 -xprefetch_auto_type=indirect_array_access -M map.4M.align |
C++ benchmarks:
| -g -fast -fma=fused -xipo=2 -xpagesize=4M -xprefetch_level=2 -xdepend -xalias_level=compatible -library=stdcxx4 -M map.4M.align |
Fortran benchmarks:
| -g -fast -fma=fused -xipo=2 -xpagesize=4M -xprefetch_level=2 -M map.4M.align |
Benchmarks using both Fortran and C:
| -g -fast(cc) -fast(f90) -fma=fused -xipo=2 -xpagesize=4M -xprefetch_level=2 -xalias_level=std -xprefetch_level=3 -xprefetch_auto_type=indirect_array_access -M map.4M.align |
Base Other Flags
C benchmarks:
| -xjobs=32 -V -# |
C++ benchmarks:
| -xjobs=32 -verbose=diags,version |
Fortran benchmarks:
| -xjobs=32 -V -v |
Benchmarks using both Fortran and C:
| -xjobs=32 -V -# -v |
Peak Compiler Invocation
C benchmarks:
| 433.milc: | cc |
| 470.lbm: | cc |
| 482.sphinx3: | cc |
C++ benchmarks (except as noted below):
| CC | |
| 450.soplex: | CC |
| 453.povray: | CC |
Fortran benchmarks (except as noted below):
| f90 | |
| 410.bwaves: | f90 |
Benchmarks using both Fortran and C:
| cc f90 |
(*) Indicates a compiler flag that was found in a non-compiler variable.
Peak Optimization Flags
C benchmarks:
C++ benchmarks:
Fortran benchmarks:
| 410.bwaves: | -g -fast -fma=fused -Wc,-Qiselect-funcalign=64 -M map.256M.align -xipo=2 -xprefetch_level=2 -xpagesize=256M -lfast |
| 416.gamess: | -g -xprofile=collect:./feedback(pass 1) -xprofile=use:./feedback(pass 2) -fast -fma=fused -Wc,-Qiselect-funcalign=64 -M map.4M.align -xipo=2 -xprefetch=no%auto -xpagesize=4M |
| 434.zeusmp: | basepeak = yes |
| 437.leslie3d: | -g -xprofile=collect:./feedback(pass 1) -xprofile=use:./feedback(pass 2) -fast -fma=fused -Wc,-Qiselect-funcalign=64 -M map.4M.align -xipo=2 -xpagesize=4M -Wc,-xthroughput -W2,-xthroughput=yes -xprefetch_level=1 |
| 459.GemsFDTD: | -g -xprofile=collect:./feedback(pass 1) -xprofile=use:./feedback(pass 2) -fast -fma=fused -Wc,-Qiselect-funcalign=64 -M map.4M.align -xipo=2 -xprefetch_level=3 -xpagesize=4M -Wc,-xthroughput -W2,-xthroughput=yes |
| 465.tonto: | -g -xprofile=collect:./feedback(pass 1) -xprofile=use:./feedback(pass 2) -fast -xpagesize=4M -stackvar -Wc,-xthroughput -W2,-xthroughput=yes -M map.4M.align -xipo=2 -lfast |
Benchmarks using both Fortran and C:
| 435.gromacs: | -g -xprofile=collect:./feedback(pass 1) -xprofile=use:./feedback(pass 2) -fast(cc) -fast(f90) -fma=fused -Wc,-Qiselect-funcalign=64 -M map.4M.align -xipo=2 -xpagesize=4M -Wc,-xthroughput -W2,-xthroughput=yes |
| 436.cactusADM: | -g -fast(cc) -fast(f90) -fma=fused -Wc,-Qiselect-funcalign=64 -xalias_level=std -xprefetch_auto_type=indirect_array_access -M map.4M.align -xipo=2 -xprefetch_level=2 -xpagesize=4M -Wc,-xthroughput -W2,-xthroughput=yes |
| 454.calculix: | -g -fast(cc) -fast(f90) -fma=fused -Wc,-Qiselect-funcalign=64 -xalias_level=std -xprefetch_auto_type=indirect_array_access -M map.4M.align -xipo=2 -xprefetch_level=2 -xprefetch=latx:3 -xpagesize=4M |
| 481.wrf: | basepeak = yes |