IBM Corporation IBM Power E870 CODE DEFECT (CD) RESULT
Tested by: IBM Corporation Test Sponsor: IBM Corporation Test location: Austin, TX Test date: Sept 24, 2014
SPEC license #: 11 Hardware Availability: Nov-2014 Software Availability: Nov-2014 Publication: Nov 06, 2014
Validation Errors Summary
SPECjbb2013 results have been marked as "Code Defect" due to a bug that impacts the comparability of results. The bug can significantly reduce the amount of work done during the measurement period, resulting in an inflated SPECjbb2013 metric.
Benchmark Results Summary
 
Overall Throughput RT curve
Overall SUT (System Under Test) Description
VendorIBM Corporation
Vendor URLhttp://www.ibm.com
System SourceSingle Supplier
System DesignationServer Rack
Total Systems1
All SUT Systems IdenticalYES
Total Nodes1
All Nodes IdenticalYES
Nodes Per System1
Total Chips8
Total Cores80
Total Threads640
Total Memory Amount (GB)4096
Total OS Images1
SW EnvironmentNon-virtual
 
Hardware hw_1
NameIBM Power E870
VendorIBM Corporation
Vendor URLhttp://www.ibm.com
AvailableNov-2014
ModelIBM Power E870
Form Factor12U
CPU NamePOWER8
CPU CharacteristicsIntelligent Energy Optimization enabled up to 4.46 GHz
Number of Systems1
Nodes Per System1
Chips Per System8
Cores Per System80
Cores Per Chip10
Threads Per System640
Threads Per Core8
VersionNone
CPU Frequency (MHz)4192
Primary Cache32 KB I + 64 KB D on chip per core
Secondary Cache512 KB I+D on chip per core
Tertiary Cache8 MB per core
Other Cache16 MB I+D off chip per CDIMM
Disk2 x 300 GB 15K SAS
File Systemxfs
Memory Amount (GB)4096
# and size of DIMM(s)64 x 64 GB CDIMMs
Memory DetailsDDR3 1600 MHz
# and type of Network Interface Cards (NICs)6x 10GBit and 24x 1GBit network ports
Power Supply Quantity and Rating (W)4 x 1925W (2 are redundant)
Other HardwareNone
Cabinet/Housing/EnclosureNone
Shared DescriptionNone
Shared CommentNone
NotesNone
Other Hardware network_1
NameNone
VendorNone
Vendor URLNone
VersionNone
AvailableNone
BitnessNone
NotesNone
Operating System os_1
NameRed Hat Enterprise Linux
VendorRed Hat
Vendor URLhttp://www.ibm.com
VersionRed Hat Enterprise Linux Server release 7.0 (Maipo)
AvailableJun-2014
Bitness64
NotesUpdated with a maintenance kernel: 3.10.0-123.6.3.el7.ppc64
Java Virtual Machine jvm_1
NameIBM J9 VM
VendorIBM Corporation
Vendor URLhttp://www.ibm.com
VersionIBM J9 VM 7.1 SR2 (build pxp6470_27sr2-20140804_05)
AvailableNov-2014
Bitness64
NotesNone
Other Software other_1
NameNone
VendorNone
Vendor URLNone
VersionNone
AvailableNone
BitnessNone
NotesNone
Hardware
OS Images os_Image_1(1)
Hardware Description hw_1
Number of Systems 1
SW Environment non-virtual
Tuning Enable Dynamic Power Saver mode
Notes None
OS Image os_Image_1
JVM Instances jvm_Ctr_1(1), jvm_Backend_1(4), jvm_TxInjector_1(4)
OS Image Description os_1
Tuning OS tuning:
  • ppc64_cpu --dscr=1
  • ulimit -n 800000
  • ulimit -l unlimited
  • ulimiit -s unlimited
  • echo 155000 > /proc/sys/vm/nr_hugepages
Notes DSCR is the Data Stream Control Register which controls the degree of aggressiveness of HW memory prefetching. The setting of 1 disables HW memory prefetching.
JVM Instance jvm_Ctr_1
Parts of Benchmark Controller
JVM Instance Description jvm_1
Command Line -Xms3g -Xmx3g -Xgcthreads8 -Xcompressedrefs -XtlhPrefetch -Xtrace:none
Tuning None
Notes None
JVM Instance jvm_Backend_1
Parts of Benchmark Backend
JVM Instance Description jvm_1
Command Line -Xlp -Xms204800m -Xmx204800m -Xmn102400m -Xgcthreads70 -XX:MaxDirectMemorySize=256m -Xgc:scvTenureAge=2,scvNoAdaptiveTenure -Xnocompressedrefs -XtlhPrefetch -Dcom.ibm.crypto.provider.doAESInHardware=true -Dcom.ibm.enableClassCaching=true -Xconcurrentlevel0 -Xaggressive -Xcodecache32M -Xtrace:none -Xloaminimum0.05
Tuning "numactl --physcpubind --membind" used to affinitize each Backend JVM to a single chip and its local memory, e.g.,
  • numactl --physcpubind=0-79 --membind=0 for the first Backend JVM
Notes None
JVM Instance jvm_TxInjector_1
Parts of Benchmark TxInjector
JVM Instance Description jvm_1
Command Line -Xlp -Xms6g -Xmx6g -Xmn3g -Xgcthreads8 -Xgc:scvTenureAge=2,scvNoAdaptiveTenure -Xnocompressedrefs -Dcom.ibm.crypto.provider.doAESInHardware=true -XtlhPrefetch -Dcom.ibm.enableClassCaching=true -Xconcurrentlevel0 -Xaggressive -Xtrace:none
Tuning "numactl --physcpubind --membind" used to affinitize each TxInjector JVM to the first core of each chip and its local memory, e.g.,
  • numactl --physcpubind=0-7 --membind=0 for the first Transaction Injector JVM
Notes None
max-jOPS = jOPS passed before the First Failure
Pass/Fail Pass Pass Fail Fail Fail
jOPS 1293622 1299150 1304679 1310207 1315735
critical-jOPS = Geomean ( jOPS @ 10000; 50000; 100000; 200000; 500000; SLAs )
Response time percentile is 99-th
SLA (us) 10000 50000 100000 200000 500000 Geomean
jOPS 80160 118858 135443 163085 218368 135665
  Percentile
  10-th 50-th 90-th 95-th 99-th 100-th
500us - / 5528 - / 5528 - / 5528 - / 5528 - / 5528 - / 5528
1000us 5528 / 11057 - / 5528 - / 5528 - / 5528 - / 5528 - / 5528
5000us 259830 / 265358 176906 / 182434 27641 / 33170 22113 / 27641 11057 / 16585 - / 5528
10000us 270887 / 276415 237717 / 243245 171377 / 176906 127151 / 132679 82924 / 77396 - / 5528
20000us 276415 / 281943 259830 / 265358 210075 / 215604 182434 / 187962 99509 / 105038 - / 5528
50000us 281943 / 287472 270887 / 276415 232189 / 237717 204547 / 210075 116094 / 121623 27641 / 5528
100000us 287472 / 293000 276415 / 281943 237717 / 243245 204547 / 210075 132679 / 138207 27641 / 5528
200000us 298528 / 304056 281943 / 287472 248773 / 254302 221132 / 226660 160321 / 165849 38698 / 5528
500000us 1127773 / 552830 298528 / 304056 276415 / 281943 265358 / 270887 215604 / 221132 132679 / 138207
1000000us 1299150 / - 1299150 / - 315113 / 320641 298528 / 304056 276415 / 281943 171377 / 154792
Probes jOPS / Total jOPS
Request Mix Accuracy
Note
(Actual % in the Mix - Expected % in the Mix) must be within:
'Main Tx' limit of +/-5.0% for the requests whose expected % in the mix is >= 10.0%
'Minor Tx' limit of +/-1.0% for the requests whose expected % in the mix is < 10.0%
There were no non-critical failures in Response Time curve building
Delay between status pings
IR/PR Accuracy
This section lists properties only set by user
Property Name Default Controller
specjbb.comm.connect.connpool.size 256 512
specjbb.comm.connect.pool.max 256 512
specjbb.comm.connect.timeouts.connect 60000 400000
specjbb.comm.connect.timeouts.read 60000 400000
specjbb.comm.connect.timeouts.write 60000 400000
specjbb.forkjoin.workers 640 250
specjbb.group.count 1 8
specjbb.heartbeat.period 10000 20000
specjbb.heartbeat.threshold 100000 200000
specjbb.txi.pergroup.count 1 1
View table in csv format
 
Level: COMPLIANCE
Check Agent Result
Check properties on compliance All PASSED
 
Level: CORRECTNESS
Check Agent Result
Compare SM and HQ Inventory All PASSED
High-bound (max attempted) is 552830 IR
High-bound (settled) is 526455 IR