Significance of Results

Oracle's Sun SPARC Enterprise M9000 server deliveredsingle-system TPC-H @3000GB world record performance andprice performance results.The Sun SPARC Enterprise M9000 server,running Oracle Database 11g Release 2 on the Oracle Solaris operating systemproves the power of Oracle/Sun solution.

  • The Sun SPARC Enterprise M9000 serverconfigured with SPARC VII processors, Sun Storage J4200 arrays plus StorageTek 6140 arrays plusSun Storage F5100 Flash Array storage,and running Oracle Solaris 10 operating system combined with Oracle 11g Release 2 achieved World Record TPC-H performance of188,229.9 QphH@3000GB for non-clustered systems.

  • The Sun SPARC Enterprise M9000 serveroutperformed the next best competitor non-clustered system, the IBM Power 595 by 20%.

  • The Sun SPARC Enterprise M9000 serveroutperformed the HP Itanium Superdome by 3.1x.

  • The Sun Storage J4200 array configuration (a total of 32 J4200arrays) in this benchmark delivered IO performance of over 21GB/sec Read and 29 GB/sec Write.

  • This TPC-H result demonstrates that the Sun SPARC Enterprise M9000server can handlethe increasingly large databases required of DSS systems. The SunSPARC Enterprise M9000 server delivered more than 18 GB/sec of real IOthroughput with Oracle 11g Release 2.

Performance Landscape

TPC-H @3000GB performance chart, non-clustered systems

CPU type processors/cores
Memory Composite
(QphH) $/perf
($/QphH) Power
(QppH) Throughput
(QthH) Database Available Sun SPARC Enterprise M9000
2.88GHz SPARC64 VII 32p/128c
512GB 188,229.9 $23.99 175,850.0 201,481.3 Oracle 04/10/10 IBM Power 595
5.0GHz POWER6 32p/64c
512GB 156,537.3 $20.60 142,790.7 171,607.4 Sybase 11/24/09 Sun Fire E25K
1.8GHz UltraSPARC IV+ 72p/144c
288GB 114,713.7 $36.68 136,798.4 96,194.3 Oracle 04/09/07 Unisys ES7000 7600R
2.6GHz Intel Xeon 16p/96c
1024GB 102,778.2 $21.05 120,254.8 87,841.4 SQL Server 05/06/10 HP Integrity Superdome
1.6GHz Intel Itanium 32p/64c
256GB 60,359.3 $32.60 80,838.3 45,068.3 SQL Server 05/21/07 QphH = the Composite Metric (bigger is better)
$/QphH = the Price/Performance metric (smaller is better)
QppH = the Power Numerical Quantity
QthH = the Throughput Numerical Quantity

Complete benchmark results found at the TPC benchmark website

Results and Configuration Summary


Sun SPARC Enterprise M9000

32 x SPARC VII 2.88 GHz processors
512 GB memory
2 x internal SAS (2 x 146 GB)

External Storage:

32 x Sun Storage J4200 arrays (32x 12 x 600 GB)
2 x StorageTek 6140 Arrays (2x 16 x 300 GB)
1 x Sun Storage F5100 Flash Array ( 80 x 24 GB)


Operating System: Oracle Solaris 10 10/09
Database: Oracle 11g Release 2 Enterprise Edition
Audited Results:

Database Size: 3000 GB (Scale Factor 3000) TPC-H Composite: 188,229.9 QphH@3000GB Price/performance: $23.99/QphH@3000GB Available: 04/10/2010 Total 3 year Cost: $4,515,692 TPC-H Power: 175,850.0 TPC-H Throughput: 201,481.3 Database Load Time: 4:43:25
Benchmark Description

The TPC-H benchmark is a performance benchmark established by theTransaction Processing Council (TPC) to demonstrate DataWarehousing/Decision Support Systems (DSS). TPC-H measurements areproduced for customers to evaluate the performance of various DSSsystems. These queries and updates are executed against a standarddatabase under controlled conditions. Performance projections andcomparisons between different TPC-H Database sizes (100GB, 300GB,1000GB, 3000GB and 10000GB) are not allowed by the TPC. TPC-H is a data warehousing-oriented, non-industry-specific benchmarkthat consists of a large number of complex queries typical of decisionsupport applications. It also includes some insert and delete activitythat is intended to simulate loading and purging data from a warehouse.TPC-H measures the combined performance of a particular databasemanager on a specific computer system.

The main performance metric reported by TPC-H is called the TPC-HComposite Query-per-Hour Performance Metric (QphH@SF, where SF is thenumber of GB of raw data, referred to as the scale factor). QphH@SF isintended to summarize the ability of the system to process queries inboth single and multi user modes. The benchmark requires reporting ofprice/performance, which is the ratio of QphH to total HW/SW cost plus3 years maintenance. A secondary metric is the storage efficiency,which is the ratio of total configured disk space in GB to the scalefactor.

Key Points and Best Practices

  • Out of box performance. There was hardly any system tuning applied.

  • Best practice storage configuration - Tables on Sun Storage J4200 (high I/O bandwidth), Indexes on Sun Storage F5100 Flash Array (low latency), Redo logs on StorageTek 6140 array (cached array).

  • The storage was managed by Oracle ASM with allocation unit 4M (default 1M). Larger allocation unit means less number of extents and suitable for large database.

  • All the Oracle files except TEMP tablespace were mirrored under Oracle ASM and IO performance was good and balanced across all the disks. Also because of the Oracle ASM mirror it passed ACID test.

  • Oracle tables were 128-way partitioned and parallel degree for each table was set to 128 because the system had 128 cores.This setting worked the best for performance.

  • CPU was 100% utilized while maintaining high IO rate with about 900 parallel processes during the Throughput run. IO peak from Oracle was over 18GB/sec during both Power and Throughput run.
See Also

Benchmark Tags

Current, World-Record, IBM, HP, $/Perf, TPC-H, Performance, Bandwidth, Solaris, SPARC64

Read More about [TPC-H@3000GB World Record Single-System Sun/Oracle...