Innovation: Setting World Records in TPC-H Paraccel & Sun Fire X4100 SF1000
Tuesday Oct 30, 2007
Innovation is an infectious thing in a very good way. Innovators seek out innovators to get even greater things. This is a story that is all about that. Customer wanted innovative hardware, they came to Sun, customer was interested in new database software (Paraccel). We saw each other's innovations, did some test, made some plans, and became quick friends. Wait till we apply a few more innovative things and make things even better.
Sun Fire X4100 using Paraccel Analytic Database sets World Record Performance and Price/Performance for the TPC-H @1000GB. This benchmark result demonstrates that a shared-nothing cluster of Sun Fire X4100s achieves the best performance and at the same time the best price/performance amongst all TPC-H@1000GB submissions. Each of the 48 node in the cluster was powered by 2 dual-core 2.8GHz Opteron processors.
The best thing:The Paraccel Analytic Database isn't about setting benchmark world records. It SETS world records because it is designed to do industrial strength everyday DSS. If you need this kind of database work for mission critical activities you should get to know Paraccel.
The benchmark details: the Sun Fire X4100 cluster achieved a QphH@1000GB of 315,842 together with $/QphH@1000GB of $4.57. The performance is almost 5 times the previously reported best performance and the price/performance is almost twice as good as the previously reported best price/performance.
Other Innovations: Another key area of innovation is business model. I've long complained about pricing a gallon of automotive gasoline by the number of cylinders in each particular car (wow that sounds silly) - whoops, I mean pricing software by # of cores. Paraccel Analytic Database innovates here as well. Databases manage data, pricing is tied to the amount of data managed. Fair and easy to measure and related to what the software does. The majors could learn from Paraccel in this area as well.
TPC-H @SF1000GB Performance Table:
$/QphH = TPC-H Price/Performance metric (smaller is better)
QphH = TPC-H Composite Metric (bigger is better)
| System | Socket /Core/ Thread |
CPU Type GHz | QphH | $/ QphH |
Total Price | DBMS |
Avail | # Clust node |
| Sun Fire X4100 |
96/ 192/ 192 |
Opt 2.8GHz |
315,842 |
4.57 |
624,770 |
Paraccel |
10/29/07 | 48 |
| HP Super- dome |
32/ 64/ 64 | Itan2 1.6GHz | 69,999 |
28.69 |
2,008,168 |
SQLS |
6/18/07 | 1 |
| HP Super- dome |
64/ 64/ 64 | Itan2 1.6GHz | 68,101 |
59.00 |
4,008,065 |
Oracle |
1/18/06 | 1 |
| IBM eSvr 346 |
64/ 64/ 128 | Xeon 3.6GHz | 53,451 |
32.80 |
1,753,144 |
DB2 |
2/14/05 | 64 |
| HP DL585 |
48/ 48/ 48 | Opt 2.2GHz |
35,141 |
59.93 |
2,106,123 |
Oracle |
10/21/04 | 12 |
| Bull 6320 |
32/ 32/ 32 | Itan2 1.6GHz | 34,988 |
38.41 |
1,343,811 |
Oracle |
2/02/06 | 1 |
TPC-H @1000GB Price-Performance Table:
$/QphH = TPC-H Price/Performance metric (smaller is better)
QphH = TPC-H Composite Metric (bigger is better)
| System | Socket/ Core/ Thread |
CPU Type GHz | QphH | $/ QphH |
Price in US dollars |
DBMS |
Available | #
Cluster Node |
| Sun Fire X4100 |
96/ 192/ 192 |
Opt 2.8GHz | 315842 |
4.57 |
624,770 |
Paraccel |
10/29/07 | 48 |
| Sun Fire X4500 G2 |
2/ 4/ 4 |
Opt 2.8 GHz | 5604 |
8.11 |
45,439 |
SybIQ |
10/15/07 | 1 |
| HP DL585 G2 |
4/ 8/ 8 |
Opt 2.8GHz | 14772 |
9.73 |
143,736 |
SQLS |
04/25/07 | 1 |
| Bull 3045 |
4/ 8/ 16 | Itan2 1.6GHz | 12087 |
12.56 |
151,870 |
SQLS |
03/06/07 |
1 |
| HP DL585 G1 |
4/ 4/ 4 | Opt 2.4GH | 10493 |
13.85 |
145,264 |
SQLS |
03/02/06 |
1 |
| Bull 5651 |
16/ 16/ 16 | Itan2 1.6GHz | 17059 |
25.48 |
434,553 |
SQLS |
05/08/06 |
1 |
| IBM eSvr 346 |
64/ 64/ 128 | Xeon 3.6GHz | 53451 |
32.80 |
1,753,144 |
DB2 |
02/14/05 |
64 |
Benchmark Description
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. A secondary metric is the storage efficiency, which is the ratio of total configured disk space in GB to the scale factor.
Disclosure Statement:
TPC-H @1000GB Sun Fire X4100 cluster 315,842.9 QphH@1000GB, $4.57/QphH@1000GB, avail 10/29/07; TPC-H, QphH, $/QphH tm of Transaction Processing Performance Council (TPC). More info http://www.tpc.org.
See Also:
- Sun Fire X4100 TPC-H Executive Summary Report Acrobat PDF (68K)
- Complete Sun Fire X4100 TPC-H Full Disclosure Report Acrobat PDF (590K)
- Transaction Processing Performance Council (TPC) Home Page
- Ideas International Benchmark page
Results Summary
| Audited Results | |||
| |
Database Size: | 1000 GB (Scale Factor 1000) | |
| |
TPC-H Composite: | 315,842 QphH@1000GB | |
| |
Price/performance: | $4.57 / QphH@1000GB | |
| |
Available | Oct 29, 2007 | |
| Number of Systems: | 48 x Sun Fire X4100, each X4100: 16GB memory each, 2 x 146GB (10k RPM) internal SAS each |
||
| Total # Procs: | 96 |
||
| Total # Cores: | 192 |
||
| Total # Threads: | 192 |
||
| Processor GHz: | Opteron/2.8 GHz Dual-core |
||
| Storage: | 13053 Total Gbytes of disk | ||
| Database: | Paraccel Analytic Database | ||
| Operating System: | RedHat Enterprise Linux 4.4 | ||
| Total 3 year Cost: | $1,442,050.52 | ||
| Other Performance Metrics | |||
| |
TPC-H Power: | 289,877.5 |
|
| |
TPC-H Throughput: | 344,134.1 |
|
| |
Database Load Time | 41 minutes 44 seconds |
|
| |
Storage Ratio: | 13.05 |
|
|
|
|
|
Total system price for ParAccel is obviously incor...