SPEC Seal of Reviewal SPECjEnterprise®2010 Result
Copyright © 2009-2012 Standard Performance Evaluation Corporation
WebSphere Application Server V8.5 on 16-core LPAR IBM Power 780 and DB2 10.1 on 16-core LPAR IBM Power 780
10,902.30 SPECjEnterprise2010 EjOPS
Submitter:
IBM Corporation
SPEC license # 11 Test date:
Sep-2012
Software Products
WebSphere Application Server V8.5
IBM J9 VM (build 2.6, JRE 1.7.0 IBM J9 AIX ppc-32)
IBM J9 VM (build 2.6, JRE 1.6.0 IBM J9 Linux x86-32)
IBM DB2 Universal JDBC Drivers (3.63.123)
DB2 10.1
IBM InfoSphere Optim pureQuery Runtime v3.1.1
Software Configurations
JEE Application Server
Emulator Software Config
Database Software Config
Driver Config
Hardware Systems
JEE AppServer & Database Server HW
Emulator HW
Load Driver HW

System Configuration Diagram

Benchmark
Modifications
Configuration
Bill of Materials
Other Info
General Notes
Full Disclosure Archive

SUT Configuration
JEE Server Nodes: 1   DB Server Nodes: 1
JEE Server CPUs: 16 cores, 4 chips   DB Server CPUs: 16 cores, 4 chips
JEE Instances: 8   DB Instances: 1
Other SUT Components:
   

Benchmark Settings
Dealer Injection Rate: 6,845   Ramp Up Time: 1,400 seconds
DB Load Injection Rate: 7,000   Steady State Time: 3,600 seconds
# of Dealer Agents: 68,450   Ramp Down Time: 300 seconds
# of Manufacturing Agents: 20,535  

Detailed Results
Dealer Transaction Mix Count Tx. Mix Failure Count
Purchase 6,173,771 25.00% 3
Manage 6,177,009 25.01% 4
Browse 12,342,644 49.98% 0
Total # Dealer Transactions 24,693,424
Manufacturing Transaction Mix Count Tx. Mix Failure Count
CreateVehicleEJB 7,275,683 49.99% 908
CreateVehicleWS 7,279,159 50.01% 1,136
Total # Manufacturing Transactions 14,554,842
Response Times Average Standard Deviation Maximum 90th% Reqd
Purchase 0.374 0.98 21.153 1.030 2.000
Manage 0.267 0.53 22.327 0.690 2.000
Browse 0.462 1.11 17.389 1.520 2.000
CreateVehicleEJB 0.050 0.06 10.321 0.100 5.000
CreateVehicleWS 0.101 0.21 11.282 0.200 5.000
Reproducibility Run (SPECjEnterprise2010 EjOPS) 10,918.27

WebSphere Application Server V8.5
Vendor: IBM Corporation Availability: Dec-2012
CTS Version: 6.0 Date Passed CTS: May-2012
OS Name: IBM AIX V7.1 TL2
EJB Protocol: RMI/IIOP

IBM J9 VM (build 2.6, JRE 1.7.0 IBM J9 AIX ppc-32)
Vendor: IBM Corporation Availability: Dec-2012
OS Name: IBM AIX V7.1 TL2

IBM J9 VM (build 2.6, JRE 1.6.0 IBM J9 Linux x86-32)
Vendor: IBM Corporation Availability: Jun-2011
OS Name: Novell SUSE Linux Enterprise Server 11 SP1

IBM DB2 Universal JDBC Drivers (3.63.123)
Vendor: IBM Corporation Availability: Apr-2012
OS Name: IBM AIX V7.1 TL2

DB2 10.1
Vendor: IBM Corporation Availability: Apr-2012
OS Name: IBM AIX V7.1 TL2

IBM InfoSphere Optim pureQuery Runtime v3.1.1
Vendor: IBM Availability: May-2012
Description: IBM InfoSphere Optim pureQuery Runtime for Linux, UNIX, and Windows Version 3.1.1

JEE Application Server
JEE Product: WebSphere Application Server V8.5
JVM Product: IBM J9 VM (build 2.6, JRE 1.7.0 IBM J9 AIX ppc-32)
JDBC Product: IBM DB2 Universal JDBC Drivers (3.63.123)
Other Product: IBM InfoSphere Optim pureQuery Runtime v3.1.1
Hardware: JEE AppServer & Database Server HW
# of Instances: 8
Interfaces:
Web Interface for the Order Domain
Web Services Interface for the Supplier Domain
EJB Interface for the Manufacturing Domain
Web Services Interface for the Manufacturing Domain
Notes / Tuning Information
JVM Parameters:
  -Xms2368m -Xmx2368m -Xmo700m -Xmn1668m  -Xnoloa
  -Xgcthreads8 -Xss128k -Xgcpolicy:gencon -Xdisableexplicitgc
  -Xshareclasses:none -Xaggressive -XtlhPrefetch -Xcodecache14m -Xtrace:none -Xlp
  -Xnodfpbd -Xnocompactgc -Xnoclassgc
  -Dcom.ibm.CORBA.FragmentSize=10000 -Djavax.xml.transform.TransformerFactory=org.apache.xalan.processor.TransformerFactoryImpl
  -Djavax.xml.parsers.SAXParserFactory=org.apache.xerces.jaxp.SAXParserFactoryImpl
  -Djavax.xml.parsers.DocumentBuilderFactory=org.apache.xerces.jaxp.DocumentBuilderFactoryImpl
  -Djava.net.preferIPv4Stack=true -Dsun.net.inetaddr.ttl=0
  -Dcom.ibm.websphere.jpa.entitymanager.poolcapacity=162
  -Dcom.ibm.websphere.ejbcontainer.excludeRootExceptionOnRollback=true
  -Dcom.ibm.websphere.webservices.http.connectionKeepAlive=true
  -Dcom.ibm.websphere.webservices.http.maxConnection=50
  -Dcom.ibm.websphere.webservices.http.connectionIdleTimeout=600
  -Dcom.ibm.websphere.webservices.http.connectionTimeout=0
  -Dcom.ibm.websphere.webservices.http.connectionPoolCleanUpTime=6000
  -DdisableWSAddressCaching=true
  -Dclient.encoding.override=ISO-8859-1
  -DAIODisableTimeouts=true

Refer to files appservers/node.resources.xml, and appservers/server.xml
for all server configuration and resource configuration information


WebSphere Application Server APAR Fixes PM61650 PM61561 PM72010 PM71704 PM71833 PM71042
InfoSphere Optim pureQuery Runtime is used by WebSphere Applicatioin Server for efficient statement batching  with APAR iFix PM71704

Emulator Software Config
JEE Product: WebSphere Application Server V8.5
JVM Product: IBM J9 VM (build 2.6, JRE 1.7.0 IBM J9 AIX ppc-32)
Hardware: Emulator HW
# of Instances: 4
Notes / Tuning Information
JVM Parameters:
  -Xms2048m -Xmx2048m -Xmo500m -Xmn1548m -Xgcpolicy:gencon
  -Xnoloa -Xgcthreads4 -Xdisableexplicitgc -Xss128k -Xtrace:none
  -Djava.net.preferIPv4Stack=true -Dsun.net.inetaddr.ttl=0
  -Djavax.xml.transform.TransformerFactory=org.apache.xalan.processor.TransformerFactoryImpl
  -Djavax.xml.parsers.SAXParserFactory=org.apache.xerces.jaxp.SAXParserFactoryImpl
  -Djavax.xml.parsers.DocumentBuilderFactory=org.apache.xerces.jaxp.DocumentBuilderFactoryImpl
  -Dcom.ibm.websphere.naming.hostname.normalizer=...none...
  -DdisableWSAddressCaching=true
  -Dcom.ibm.websphere.webservices.http.connectionKeepAlive=true
  -Dcom.ibm.websphere.webservices.http.maxConnection=50
  -Dcom.ibm.websphere.webservices.http.connectionIdleTimeout=600
  -Dcom.ibm.websphere.webservices.http.connectionTimeout=0

Refer to files emulator/server.xml
for all server configuration and resource configuration information


Database Software Config
DB Product: DB2 10.1
Hardware: JEE AppServer & Database Server HW
# of Instances: 1
Character Set: UTF-8
Notes / Tuning Information
DB2_COLLECT_TS_REC_INFO=FALSE
DB2_APM_PERFORMANCE=ON
DB2_KEEPTABLELOCK=CONNECTION
DB2_USE_ALTERNATE_PAGE_CLEANING=ON
DB2_MINIMIZE_LISTPREFETCH=YES
DB2_RESOURCE_POLICY=/home/db2inst1/scripts/loggerpri.cfg
DIAGLEVEL 3
DFT_MON_BUFPOOL off
DFT_MON_LOCK off
DFT_MON_SORT off
DFT_MON_STMT off
DFT_MON_TIMESTAMP off
DFT_MON_UOW off
DFT_MON_TABLE off
MAX_COORDAGENTS 350 automatic
CPUSPEED -1
HEALTH_MON off
SELF_TUNING_MEM off
LOGBUFSZ 4096
LOGSECOND 0
LOGPRIMARY 150
SOFTMAX 3000
CHNGPGS_THRESH 99
LOGFILSIZ 65535
NUM_IOCLEANERS 16
AUTO_MAINT off
AUTO_RUNSTATS off
AUTO_TBL_MAINT off
LOCKLIST 100000
MAXLOCKS 100
MINCOMMIT 1
MON_REQ_METRICS none
MON_ACT_METRICS none
MON_OBJ_METRICS none
AUTO_STMT_STATS off
AUTO_MAINT off
AUTO_RUNSTATS off
AUTO_TBL_MAINT off
O_ITEM  volatile
S_SUPPLIER volatile
M_WORKORDER volatile
O_ORDERLINE append on
O_CUSTINVENTORY append on
O_ORDERS append on
S_PURCH_ORDER append on
S_PURCH_ORDERLINE append on
M_INVENTORY volatile
DB2COMM=TCPIP
IBMDEFAULTBP immediate size 12000000
SPEC_BP immediate size 2000
JMS_BP immediate size 4000
M_INV_BP immediate size 500000
M_PARTS_BP immediate size 120000
M_BOM_BP immediate size 60000
M_WO_BP immediate size 20000
M_WO2_BP immediate size 2000
M_OL_BP immediate size 2000
O_O_BP immediate size 140000
O_O2_BP immediate size 60000
O_OL_BP immediate size 100000
O_CI_BP immediate size 400000
O_C_BP immediate size 200000
O_ITEM_BP immediate size 2000
O_ITEM2_BP immediate size 2000
S_C_BP immediate size 200000
S_S_C_BP immediate size 400000
S_S_BP immediate size 2000
S_P_O_BP immediate size 2000
S_P_O2_BP immediate size 2000
S_P_OL_BP immediate size 3000
O_ORDERS_BP immediate size 60000
O_CUSTOMER_BP immediate size 90000
S_COMPONENT_BP immediate size 90000
INDEX_BP immediate size 200000
DATA_BP immediate size 140000

Driver Config
JVM Product: IBM J9 VM (build 2.6, JRE 1.6.0 IBM J9 Linux x86-32)
Hardware: Load Driver HW
# of Instances: 4
Notes / Tuning Information
JVM Parameters:
  -Xms2048M -Xmx2048M
  -Djava.net.preferIPv4Stack=true
  -Dsun.net.inetaddr.ttl=0
  -Dnetworkaddress.cache.ttl=0
  -Dcom.ibm.websphere.naming.jndicache.cacheobject=none
  -Dcom.ibm.websphere.naming.hostname.normalizer=...none...
  -DdisableWSAddressCaching=true
  -Dcom.ibm.CORBA.ConnectionMultiplicity=20
  -Dcom.ibm.websphere.webservices.http.connectionKeepAlive=true
  -Dcom.ibm.websphere.webservices.http.maxConnection=810
  -Dcom.ibm.websphere.webservices.http.connectionIdleTimeout=600
  -Dcom.ibm.websphere.webservices.http.connectionTimeout=0

JEE AppServer & Database Server HW (SUT hardware)
Hardware Vendor: IBM Corporation OS Vendor: IBM
Model Name: IBM Power 780 OS Name: IBM AIX V7.1 TL2
Processor: POWER7+ (Intelligent Energy Optimization enabled up to 4.48 GHz) Filesystem: AIX/JFS2
MHz: 4424 Disks: 4x146GB 15K SAS
# of CPUs: 32 cores, 8 chips, 4 cores/chip (16 HW threads/chip) Network Interface: 12x 10GBit network ports
Memory (MB): 786432 Other Hardware: 2xDS5324 Disk System with 48x146GB 15K E-DDM Disks, 1x v7000 Disk Controller with 16x 400GB SSD Disks
DIMM[0] Count: 16 DIMM[0] Size: 16384
DIMM[1] Count: 16 DIMM[1] Size: 32768
L1 Cache: 32 KB I + 32 KB D on chip per core # of Systems: 1
L2 Cache: 256 KB I+D on chip per core H/W Available: Oct-2012
Other Cache: 10 MB I+D on chip per core OS Available: Oct-2012
Notes / Tuning Information
System Configuration:
Two 16-core logical partitions (LPAR) were hosted on a 32-core IBM Power 780.
The first 16-core LPAR for the WebSphere Application Server was configured with
  16 dedicated cores
  256 GB of memory
  6 dual port 10 Gbit Network adapters  (with a single port being used on two of them).

The second 16-core LPAR for the DB2 Server was configured with
  16 dedicated cores
  512 GB of memory
  2 single port 10 Gbit Network adapters.
  1 embedded 10 Gbit network port
  3 quad port 8 Gb FC adapters  (with two ports being used on each of them).
The full details of the LPAR configuration are in the  lpar_cfg.pdf included in the FDA.

Tunings:
vmo -o lgpg_regions=16384 -o lgpg_size=16777216 -o v_pinshm=1
dscrctl -n -s 1
MEMORY_AFFINITY=MCM
EXTSHM=OFF
SPINLOOPTIME=4000
AIXTHREAD_SCOPE=S
LDR_CNTRL=DATAPSIZE=64k@STACKPSIZE=64k@TEXTPSIZE=64k@MAXDATA=0xB0000000@DSA
ulimit -s unlimited; ulimit -d unlimited; ulimit -m unlimited; ulimit -f unlimited; ulimit -n unlimited
ctctrl -P memtraceoff
errctrl -P errcheckoff
raso -r -o mtrc_enabled=0
skeyctl -u 0
skeyctl -k off
no -r -o net_malloc_police=0
setsched -P 61 -S rr -p $$
asoo -o aso_active=0
Network Tuning:
  rxbuf_pool_sz=4096
  rxdesc_que_sz=2048
  rx_coalese=64
  intr_rate=2500
  Network interrupt bound to a core corresponding to the server instance

Server instances were started using execrset, binding 2 instances per chip.


Emulator HW (non-SUT hardware)
Hardware Vendor: IBM Corporation OS Vendor: IBM Corporation
Model Name: IBM Power 730 Express (3.55 GHz, 16 core) OS Name: IBM AIX V7.1 TL2
Processor: POWER7 Filesystem: AIX/JFS2
MHz: 3556 Disks: 1x146GB 15K RPM
# of CPUs: 16 cores, 2 chips, 8 cores/chip Network Interface: 2x 10GBit
Memory (MB): 131072 Other Hardware:
DIMM[0] Count: 16 DIMM[0] Size: 8192
L1 Cache: 32 KB I + 32 KB D on chip per core # of Systems: 1
L2 Cache: 256 KB I+D on chip per core H/W Available: Sep-2010
Other Cache: 4 MB I+D on chip per core OS Available: Sep-2010
Notes / Tuning Information
vmo -o lgpg_regions=22400 -o lgpg_size=16777216 -o v_pinshm=1
ulimit -s unlimited; ulimit -d unlimited; ulimit -m unlimited; ulimit -f unlimited; ulimit -n unlimited
dscrctl -n -s 1
MEMORY_AFFINITY=MCM
LDR_CNTRL=DATAPSIZE=64k@STACKPSIZE=64k@TEXTPSIZE=64k
ctctrl -P memtraceoff
errctrl -P errcheckoff
raso -r -o mtrc_enabled=0
skeyctl -u 0
skeyctl -k off
no -r -o net_malloc_police=0
Queue depth set to 32 for all external storage disks

Load Driver HW (non-SUT hardware)
Hardware Vendor: IBM Corporation OS Vendor: Novell
Model Name: IBM BladeCenter HS22 OS Name: Novell SUSE Linux Enterprise Server 11 SP1
Processor: Xeon X5670 Filesystem: ext3
MHz: 2933 Disks: 2 x 146GB 15K SAS
# of CPUs: 12 cores, 2 chips, 6 cores/chip Network Interface: 1x 10GBit and 1x 1GBit network ports
Memory (MB): 49152 Other Hardware:
L1 Cache: 32KB(I)+32KB(D) # of Systems: 8
L2 Cache: 256 I+D on chip per core H/W Available: Mar-2010
Other Cache: 12 MB I+D on chip per chip OS Available: Jun-2010
Notes / Tuning Information
fs.file-max = 1048576
kernel.sem = 250 32000 100 128
kernel.shmall = 4294967296
kernel.shmmax = 68719476736
net.core.netdev_max_backlog = 400000
net.core.optmem_max = 30000000
net.core.rmem_default = 30000000
net.core.wmem_default = 30000000
net.core.rmem_max = 30000000
net.core.wmem_max = 30000000
net.core.somaxconn = 300000
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.ip_forward = 0
net.ipv4.ip_local_port_range = 1024 65000
net.ipv4.tcp_fin_timeout = 10
net.ipv4.tcp_max_syn_backlog = 30000
net.ipv4.tcp_max_tw_buckets = 2000000
net.ipv4.tcp_mem  = 30000000 30000000 30000000
net.ipv4.tcp_rmem = 30000000 30000000 30000000
net.ipv4.tcp_wmem = 30000000 30000000 30000000
net.ipv4.tcp_timestamps = 0
net.ipv4.tcp_sack = 0
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.all.arp_ignore = 1
ulimit -n 1048576

Network Tuning:
  set txqueuelen 60000

Benchmark Modifications
Schema Modifications:
  Scripts to create the database are included in the FDA.
  All tables except the following are range partitioned: S_SUPPLIER, S_RUN_CONFIG, U_SEQUENCES, U_DATASEGMENT, U_LOADER, U_LOAD_MESSAGE, U_LOAD_SECTION
Load Program Modifications:
  No modifications to the load programs were made.

Benchmark Configuration Information
Isolation Requirement Info:
  The benchmark meets the isolation level requirements by version column checking on entities against the database.
  The ItemEnt bean was cached for 20 minute intervals using the WebSphere JPA Object Cache mechanism.
Durability Requirement Info:
  To ensure database durability, RAID10 was used for the
    database logs maintained on the external storage.
    A total of 48 x 146GB disks were used for the log.
Storage Requirement Info:
  Over the course of a 88 minute run at an injection rate of
    6845, the database storage for database system increased by
    7 GB.  Given a linear scale, a 24 hour run at the same
    injection rate would increase storage for database system
    by 115 GB.  One RAID10 disk array was used to create the
    filesystem space for the database.  The array contained
    16 x 400GB disks providing 2 TB of storage.

Bill of Materials
Supplier  Description                                         Product #           Qty
--------  -------------------------------------------         ----------------    ---
IBM       IBM Power 780                                       9179-MHB              1
-            8x4-core POWER7+ Processor
-            (4.424GHz 10MB L3 Cache per core)
-            768GB RAM (16x8GB DIMMS, 16x16GB DIMMS),
-            5x146.8GB 15k SAS HDD
-            10 GB Ethernet SR PCI dual port                  5721                  6
-            10 GB Ethernet SR PCI                            5769                  2
-            PCIE2 8GB 4port Fibre Channel                    5729                  3
-            3 Year Onsite Repair 24x7 4 Hour Response
IBM       AIX Standard Edition V7.1 TL2 per processor         5765-G98-0017        32
IBM       AIX 3 year support per processor                    5773-SM3-1259        32
...
IBM       IBM System Storage DS3524 Express                    1746-C4A             2
IBM       146GB 15,000 rpm 6Gb SAS 2.5; HDD                    5205                48
IBM       3 Year Onsite Repair 24x7 4 Hour Response            67567DT
IBM       IBM Storwize V7000 Disk System                       5639-VM1             1
IBM       400 GB 2.5INCH SSD (E MLC)                           3543                16
IBM       3 Year Onsite Repair 24x7 4 Hour Response            67567DT
...
IBM       DB2 Enterprise Server Edition 10.1                   D597RLL             16*
-           Lic+SW Maint 12 Months
IBM       DB2 Enterprise Server Edition 10.1                   E00BILL             32
-           SW Maint Rewl 1 Anniv
IBM       WebSphere Application Server V8.5                    D55W8LL             16*
-           Lic+SW Maint 12 Months
IBM       WebSphere Application Server V8.5                    E1CBBLL             32
-           SW Maint Rewl 1 Anniv
IBM       IBM Optim Purequery Runtime for LUW 3.1.1            D61YKLL             16*
-           Lic+SW Maint 12 Months
IBM       IBM Optim Purequery Runtime for LUW 3.1.1            E048YLL             32
-           SW Maint Rewl 1 Anniv
...
* Note:  Pricing is based on Processor Value Units (PVU).  Each POWER7 core is 120 PVU.


Other Benchmark Information
  DNS round robin load balancing was used and hosted on the database machine
  All network connections were 10Gbps

General Notes
All exceptions in driver logs are from Optimistic Concurrency Checking.
These exceptions are expected in the benchmark.

The IBM DB2 Universal JDBC Drivers is a type 4 driver.

DB2 uses "Soft" checkpoint to ensure that no updates remain unflushed for
longer than the allowed time.

When DB2 UDB Server changes a database table with an update, insert, or delete
operation, the change is initially made in memory, not on disk. When there is
not enough space in the memory buffer to read in or write additional data pages,
DB2 UDB Server will make space by flushing some modified pages to disk.
Modified pages are also written to disk as part of the "Soft" checkpoint to
ensure that no updates remain unflushed for longer than the allowed time.
Before a change is made to the database, it is first recorded in the
transaction log. This ensures that the database can be recovered completely
in the event of a failure. Using the transaction log, transactions that
started but did not complete prior to a failure can be undone, and
transactions recorded as complete in the transaction log but not yet written
to disk can be redone.

DB2 UDB uses a write-ahead-logging protocol to guarantee recovery. This
protocol uses "Soft" checkpoint to write least-recently-used database pages
to disk independent of transaction commit. However, enough log information to
redo/undo the change to a database pages is committed to disk before the
database page itself is written. This protocol therefore renders checkpoint
unnecessary for DB2 UDB. For a more detailed description of the general
principles of the write-ahead-logging protocol, see the IBM research paper,
ARIES: A Transaction Recovery Method Supporting Fine Granularity Locking and
Partial Rollbacks Using Write-Ahead Logging," by C. Mohan, Database Technology
Institute, IBM Almaden Research Center. (http:// portal.acm.org/citation.cfm
?id=128770&coll=portal&dl=ACM&CFID=10343790&CFTOKEN=42047146)


For questions about this result, please contact the submitter: IBM Corporation
For other inquiries, please contact webmaster@spec.org

Copyright © 2009-2012 Standard Performance Evaluation Corporation

First published at SPEC.org on 03-Oct-2012