POWER6 err, try >10GB/s peak

Wednesday Jun 27, 2007

I finally saw an IBM power6 presentation that says:

"Balanced design with highest system bandwidth
– 2X Memory Bandwidth (> 10 GB/sec)

What happened to the implied IBM press release of 300GB/s bandwidth that could download iTune library in 30 sec... oh, that was meaningless marketing fluff(is that the right word?).

So to move the iTunes library from power6 memory to POWER6 CPU at the 10GB/s listed above, it would really take 15 hours at p570 peak memory speed(not-measured!).

postscript: Another doc listed 45GB/s SMP+IO bandwidth (whatever that means), time for IBM to publish the measured STREAM bcopy performance so we can put an end to this nonsense.

On a Sun Fire E25K going to disk(Yes IO, which is slower than IBM's memory) it would take half that 15 hour time. Sun demonstrated a delivered 21 GB/sec of delivered disk to CPU bandwidth. (yes I need to repeat 'delivered' twice as IBM has a tendency to only mention peak numbers and then omits the word 'peak'.

OK we don't have IBM delivered IO performance measured on an IBM p595, as IBM doesn't trust to share those numbers with the public...

For news article about bigger systems see this:
http://www.betanews.com/article/Sun_We_Can_Build_a_Faster_Supercomputer_Than_IBM/1182889189

JavaOne: reposting World Record Single JVM result

Thursday May 10, 2007

Sun's single-JVM World Record (table now shown ordered by bops/jvm):

• The 72-way Sun Fire E25K with 1.8 GHz US-IV+ set a World Record with a single-instance result of 1,149,100 SPECjbb2005 bops, 1,149,100 SPECjbb2005 bops/JVM on SPECjbb2005.
• The Sun Fire E25K with dual-core US-IV+ 1.8GHz and running Java SE 1.6.0_02 outperformed Fujitsu PrimeQuest 580 with dual-core Itanium 2 processors and running BEA JRockit by 4%.
• The 72-way Sun Fire E25K with 1.8 GHz US-IV+ processors set a World Record with a single-instance result of 1,149,100 SPECjbb2005 bops, 1,149,100 SPECjbb2005 bops/JVM on the SPECjbb2005 benchmark, Sun has a much higher result SPECjbb2005 using multiple JVMs running on a 1.95GHz US-IV+ processor. The Sun Fire E25K with 1.95GHz US-IV+ processors set a World Record for systems with 72 or fewer chips, achieving 2,105,264 SPECjbb2005 bops and 29,240 SPECjbb2005 bops/JVM on SPECjbb2005.

SPECjbb2005 Performance (ordered by performance bops/JVM : SPECjbb2005 Business Operations per Second, bigger is better)

System	Date	Processors		Performance
		(Chips, Cores, Threads)	GHz Type	bops	JVMs	bops/JVM
Sun Fire E25K	5/07	72, 144, 144	1.8 US-IV+	1,149,100	1	1,149,100
Fujitsu PRIMEQUEST 580	2/07	32, 64, 64	1.6 Itanium2	1,105,465	1	1,105,465
HP Superdome	9/06	64, 128, 128	1.6 Itanium 2	2,054,864	32	64,215
IBM p5 570	1/06	8, 16, 32	2.2 POWER5+	326,651	8	40,831
Fujitsu PP2500	3/06	128, 128, 128	2.08 SPARC64 V	1,251,024	32	39,095
Sun Fire E25K	5/07	72, 144, 144	1.95 US-IV+	2,105,264	72	29,240

Sun results have been submitted to SPEC for review and are on track for publication.

Benchmark Description

SPECjbb2005 (Java Business Benchmark) measures the performance of a Java implemented application tier (server-side Java). The benchmark is based on the order processing in a wholesale supplier application. The performance of the user tier and the database tier are not measured in this test. The metrics given are number of SPECjbb2005 bops (Business Operations per Second) and SPECjbb2005 bops/JVM (bops per JVM instance).

Disclosure Statement:

SPECjbb2005 Sun Fire E25K (72 chips, 144 cores, 1.8 GHz) 1,149,100 SPECjbb2005 bops, 1,149,100 SPECjbb2005 bops/JVM submitted for review; Fujitsu PrimeQuest 580 (32 chips, 64 cores, 1.6 GHz) 1,105,465 SPECjbb2005 bops, 1,105,564 SPECjbb2005 bops/JVM. SPECjbb2005 Sun Fire E25K (72 chips, 144 cores, 1.95 GHz) 2,105,264 SPECjbb2005 bops, 29,240 SPECjbb2005 bops/JVM submitted for review; Sun Fire E25K (72 chips, 144 cores, 1.95 GHz) 1,657,274 SPECjbb2005 bops, 23,018 SPECjbb2005 bops/JVM; HP Itanium Superdome (64 chips, 128 cores, 1.6 GHz) 2,054,864 SPECjbb2005 bops, 64,215 SPECjbb2005 bops/JVM; Fujitsu PRIMEPOWER 2500 (128 chips, 128 cores, 2.08 Ghz) 1,251,024 SPECjbb2005 bops, 39,095 SPECjbb2005 bops/JVM; IBM eServer p5 570 (8 chips, 16 cores, 2.2 GHz) 326,651 SPECjbb2005 bops, 40,831 SPECjbb2005 bops/JVM. SPEC, SPECjbb reg tm of Standard Performance Evaluation Corporation. Results as of 5/8/07 on www.spec.org

Certified Results
	Performance:		1,105,564 SPECjbb2005 bops
			1,105,564 SPECjbb2005 bops/JVM
	Reference Date:		May 8, 2007
Systems:			Sun Fire E25K
Processor/GHz:			72 US-IV+ 1.8 GHz
Operating System:			Solaris 10
JVM:			Java HotSpot(TM) 32-Bit Server, Version 6.0_02

SAS Extract, Transform, and Load Sun Fire E25K UltraSPARC IV+ 1.95 GHz

Wednesday Apr 18, 2007

The Sun Fire E25K running Solaris 10 11/06 and configured with Sun StorageTek 6140 arrays utilizing Sun StorageTek QFS 4.5 achieved multiple World Records on the SAS Extract, Transform, and Load (ETL) benchmarks. The EEC Enterprise Data Integration Test Suite is an application that performs large scale data integration operations for data warehousing.

A combination of a Sun E25K (72 1.95 GHz US-IV+) and 20 ST6140 storage arrays achieved World Record throughput of 5.9 TB per hour for the Bulkload with Data Validation into Text.

The Sun Fire E25K (32 1.95 GHz US-IV+) delivered a throughput of 3.02 TB/hr, which is 61% faster than recently published results by HP on a similar test using an Integrity Superdome server (64 1.6 GHz Intel Itanium2). When loading into a relational data store instead of text, the 32-way Sun Fire E25K was 74% faster than the HP Integrity Superdome on a similar test.

SAS also raised the data integration benchmark standard by increasing the workload complexity and loaded the same data into a full star schema data model in a relational data store while performing data validation, integrity constraints, dimension table builds, dimension lookups, and index generation.

Primary data transformations used in this test were Lookup, SQL Join, File Reader, Loop, User Written Code and Table Loader.

During this significantly complex task SAS sustained a data load rate of 1.93 TB/hour on the 72 processor Sun server configuration. This is also a new World Record for this level of complexity and data volume.

The Sun Fire E25K (72 1.95 GHz US-IV+) and 20 Sun StorageTek ST6140 arrays showed a performance improvement of 7% for the Bulk Load with Data Validation to Text over the previous Sun Fire E25K (72 1.8GHz US-IV+) and 20 Sun StorageTek ST6140 arrays.

The Sun Fire E25K (72 1.95 GHz US-IV+) and 20 Sun StorageTek ST6140 arrays showed a performance improvement of 5% for the Bulkload to Text over the previous Sun Fire E25K (72 1.8GHz US-IV+ ) and 20 Sun StorageTek ST6140 arrays.

The benchmark tests also highlighted new technology from SAS, including SAS Data Integration Studio 3.4, SAS Scalable Performance Data Server, SAS Grid Server, and SAS 9.13, as well as the effectiveness of SAS on the Sun QFS file system.

When performance and execution matter - SAS chooses Sun: These benchmark results represent significant engineering effort, collaboration and coordination between SAS and Sun. The results also illustrate the commitment of the two companies to provide the best solutions for the most demanding data integration requirements.

Performance Comparison

Load to Dataset

1 Million Customer Table
System	Processors		Bulk load	Full Star Schema Build with Data Validation	Bulkload with Data Validation
	Type & GHz	Chips, Cores
Sun Fire E25K	US-IV+ 1.95	72,144	5.97 TB/hr		5.90 TB/hr
Sun Fire E25K	US-IV+ 1.95	32,64			3.02 TB/hr
Sun Fire E25K	US-IV+ 1.8	72,144	5.68 TB/hr		5.53 TB/hr
Sun Fire E25K	US-IV+ 1.8	32,64		1.54 TB/hr	2.80 TB/hr
HP Integrity Superdome	Itanium2 1.6	64,64			1.88 TB/hr
Sun Fire E25K	US-IV+ 1.5	48,96	3.9 TB/hr	1.5 TB/hr	3.0 TB/hr

Load to Relational Data Store

1 Million Customer Table
System	Processors		Bulk load	Full Star Schema Build with Data Validation +	Bulkload with Data Validation	Full Star Schema Build with Data Validation ++
	Type & GHz	Chips, Cores
Sun Fire E25K	US-IV+ 1.95	72,144			4.44 TB/hr	1.93 TB/hr
Sun Fire E25K	US-IV+ 1.95	32,64			2.32 TB/hr
Sun Fire E25K	US-IV+ 1.8	72,144	4.23 TB/hr	2.64 TB/hr	4.15 TB/hr	1.86 TB/hr
Sun Fire E25K	US-IV+ 1.8	32,64		1.31 TB/hr	2.14 TB/hr
HP Integrity Superdome	Itanium2 1.6	64,64			1.33 TB/hr
100 Million Customer Table
Sun Fire E25K	US-IV+ 1.95	24,48				590 GB/hr

+     PLUS integrity constraints, dimension table builds, and dimension lookups

++   PLUS integrity constraints, dimension table builds, dimension lookups, and index generation

Benchmark Description

The ETL (Extract, Transform and Load) benchmark reads in a multi terabyte data set, performs data transformation, and loads it into SAS Intelligent Storage (datasets or SPDS), simulating an operation typical of large scale data integration operations for data warehousing. Complexity can be increased by adding in data validation, Star Schema Builds (with dimension table builds and lookups), as well as index creation.

Disclosure Statement:

SAS ETL Sun Fire E25K 5.90 TB/hr, 72 1.95 GHz US-IV+, Sun StorageTek 6140 Array, Solaris 10 11/06, Sun StorageTek QFS 4.5, SAS Enterprise Data Integration Server 9.1.3 SP4. Results as of 04/09/2007. More info www.sas.com.

Press Releases

SAS Press Release on sas.com

SAS Press Release on businesswire.com

Informatica/HP Press Release

Results Summary

	Performance:		5.97 TB/hr Bulkload to datasets
			5.90 TB/hr Bulkload to datasets with Data Validation
			4.44 TB/hr Bulkload to Relational Data Store with Data Validation
			1.93 TB/hr Full Star Schema Build to Relational Data Store with Data Validation and Index Generation
			* TB/hr metric is derived from the input data size, not the total IO bandwidth being used, which is over 2x larger the reported TB/hr result.
	Server:		Sun Fire E25K, 72 1.95 GHz US-IV+, 288GB memory.
	Storage:		20 Sun StorageTek 6140s, 146GB 15K RPM, 16 Drives per array, 320 total spindles
	Operating system:		Solaris 10 11/06
	SAS S/W:		SAS Enterprise Data Integration Server 9.1.3 SP4
	Filesystem:		Sun StorageTek QFS 4.5
Processors:			72 UltraSPARC IV+ 1.95 GHz

Like this post? | | | |

Posted at 05:00AM Apr 18, 2007 by BM Seer in Benchmarks
Tags: e25k iv+ ultrasparc us-iv+

Funny math to the core

Monday Apr 09, 2007

IBM thinks it is about the core count or performance per core. Get real. It is about the whole system. You can do the math based on the info in the TPC-H submissions below...
Sun: $4,207,126 /144 core = ?
IBM: $5,358,874 /64 core = ?

It is clear to see that IBM's cores each cost more than 2.5 times more than Sun's cores. Before you get too confused with 'rotten-to-the-core-math', just remember this. The IBM system costs more and the IBM system is a slower on the TPC-H benchmark. http://blogs.sun.com/bmseer/entry/database_world_record_sun_us.

• The Sun Fire E25K 1.8GHz outperformed the IBM p5-595 (Power5+) by 14% and also had 31% better price/performance. Also beat the p595 by 26% on the multi-user test (Throughput).
• The Sun Fire E25K beat the HP Integrity Superdome (Itanium2) by 60% on performance and 34% on price/performance. Sun also beat the Itanium2 Superdome by 72% for the multi-user test (Throughput).
• Last week Sun announced Sun Fire E25K systems with 1.95GHz processors.

TPC-H Disclosure Statement:

Sun Fire E25K 114,713.7 QphH@3000GB, $36.68/QphH@3000GB, avail 04/09/07, HP BladeSystem ProLiant BL25p cluster 64p DC 110,576.5 QohH@3000GB, $37.80/QphH@3000GB avail 06/08/06, Sun Fire E25K 105430.9 QphH@3000GB, $54.87/QphH@3000GB, avail 01/23/06, IBM eServer p5 595 100,512.3 QphH@3000GB, $53.32/QphH@3000GB, avail 03/01/06, HP Integrity Superdome 71,847.8 QphH@3000GB, $55.79/QphH@3000GB, avail 01/18/06, Sun Fire E25K 59,435.7 QphH@3000GB, $100.66/QphH@3000GB, avail 07/27/05, TPC-H, QphH, $/QphH tm of Transaction Processing Performance Council (TPC). More info www.tpc.org.