SPARC T4-4 Beats IBM POWER7 and HP Itanium on TPC-H @1000GB Benchmark
By Brian on Oct 03, 2011
Oracle's SPARC T4-4 server configured with SPARC-T4 processors, Oracle's Sun Storage F5100 Flash Array storage, Oracle Solaris, and Oracle Database 11g Release 2 achieved a TPC-H benchmark performance result of 201,487 QphH@1000GB with price/performance of $4.60/QphH@1000GB.
The SPARC T4-4 server benchmark results demonstrate a complete solution of building Decision Support Systems including data loading, business questions and refreshing data. Each phase usually has a time constraint and the SPARC T4-4 server shows superior performance during each phase.
The SPARC T4-4 server is 22% faster than the 8-socket IBM POWER7 server with the same number of cores. The SPARC T4-4 server has over twice the performance per socket compared to the IBM POWER7 server.
The SPARC T4-4 server achieves 33% better price/performance than the IBM POWER7 server.
The SPARC T4-4 server is up to 4 times faster than the IBM POWER7 server for the Refresh Function.
The SPARC T4-4 server is 44% faster than the HP Superdome 2 server. The SPARC T4-4 server has 5.7x the performance per socket of the HP Superdome 2 server.
The SPARC T4-4 server is 62% better on price/performance than the HP Itanium server.
The SPARC T4-4 server is up to 3.7 times faster than the HP Itanium server for the Refresh Function.
The SPARC T4-4 server delivers nearly the same performance as Oracle's SPARC Enterprise M8000 server, but with 52% better price/performance on the TPC-H @1000GB benchmark.
Oracle used Storage Redundancy Level 3 as defined by the TPC-H 2.14.2 specification which is the strictest level.
This TPC-H result demonstrates that the SPARC T4-4 server can deliver the performance while running the increasingly larger databases required of DSS systems. The server measured more than 16 GB/sec of IO throughput through Oracle Database 11g Release 2 software while maintaining the high cpu load.
The table below lists published non-cluster results from comparable enterprise class systems from Oracle, IBM and HP. Each system was configured with 512 GB of memory.
|SPARC Enterprise M8000
3 GHz SPARC64 VII+
16 / 64 / 128
3 GHz SPARC-T4
4 / 32 / 256
|IBM Power 780
4.14 GHz POWER7
8 / 32 / 128
|HP Superdome 2
1.73 GHz Intel Itanium 9350
16 / 64 / 64
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 http://www.tpc.org.
Configuration Summary and Results
512 GB memory
8 x internal SAS (8 x 300 GB) disk drives
Oracle Database 11g Release 2 Enterprise Edition
|Database Size:||1000 GB (Scale Factor 1000)|
|TPC-H Composite:||201,487 QphH@1000GB|
|Total 3 Year Cost:||$925,525|
|Database Load Time:||1:22:39|
The TPC-H benchmark is a performance benchmark established by the Transaction Processing Council (TPC) to demonstrate Data Warehousing/Decision Support Systems (DSS). TPC-H measurements are produced for customers to evaluate the performance of various DSS systems. These queries and updates are executed against a standard database under controlled conditions. Performance projections and comparisons 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 benchmark that consists of a large number of complex queries typical of decision support applications. It also includes some insert and delete activity that is intended to simulate loading and purging data from a warehouse. TPC-H measures the combined performance of a particular database manager on a specific computer system.
The main performance metric reported by TPC-H is called the TPC-H Composite Query-per-Hour Performance Metric (QphH@SF, where SF is the number of GB of raw data, referred to as the scale factor). QphH@SF is intended to summarize the ability of the system to process queries in both single and multi user modes. The benchmark requires reporting of price/performance, which is the ratio of QphH to total HW/SW cost plus 3 years maintenance.
Key Points and Best Practices
- Four Sun Storage F5100 Flash Array devices were used for the
benchmark. Each F5100 device contains 80 flash modules (FMODs).
FMODs from each F5100 device were connected to a single SAS 6 Gb HBA. A single
F5100 device showed 4.16 GB/sec for sequential read and demonstrated linear
scaling of 16.62 GB/sec with 4 x F5100 devices.
- The IO rate from the Oracle database was over 16 GB/sec.
- Oracle Solaris 10 8/11 required very little system tuning.
- The SPARC T4-4 server and Oracle Solaris efficiently
managed the system load of over one thousand Oracle parallel
- The Oracle database files for tables and indexes were managed by Oracle Automatic Storage Manager (ASM) with 4M stripe. Two F5100 devices were mirrored to another 2 F5100 devices under ASM. IO performance was high and balanced across all the FMODs.
- The Oracle redo log files were mirrored across the F5100 devices using Oracle Solaris Volume Manager with 128K stripe.
- Parallel degree on tables and indexes was set to 128. This setting worked the best for performance.
- TPC-H Refresh Function simulates periodical Refresh portion of Data Warehouse by adding new sales and deleting old sales data. Parallel DML (parallel insert and delete in this case) and database log performance are a key for this function and the SPARC T4-4 server outperformed both HP Superdome 2 and IBM POWER7 servers.
- Transaction Processing Performance Council (TPC) Home Page
- Ideas International Benchmark Page
- SPARC T4-4 Server
Oracle Database 11g Release 2 Enterprise Edition
Sun Storage F5100 Flash Array
TPC-H, QphH, $/QphH are trademarks of Transaction Processing Performance Council (TPC). For more information, see www.tpc.org. SPARC T4-4 201,487 QphH@1000GB, $4.60/QphH@1000GB, avail 10/30/2011, 4 processors, 32 cores, 256 threads; SPARC Enterprise M8000 209,533.6 QphH@1000GB, $9.53/QphH@1000GB, avail 09/22/11, 16 processors, 64 cores, 128 threads; IBM Power 780 QphH@1000GB, 164,747.2 QphH@1000GB, $6.85/QphH@1000GB, avail 03/31/11, 8 processors, 32 cores, 128 threads; HP Integrity Superdome 2 140,181.1 QphH@1000GB, $12.15/QphH@1000GB avail 10/20/10, 16 processors, 64, cores, 64 threads.