HP Continues Performance Leadership with Next Generation of Itanium 2 Processor; Benchmark Results Demonstrate Strength of Next-generation Itanium 2 Processor and HP Systems.

Business Editors/High-Tech Writers

PALO ALTO, Calif.--(BUSINESS WIRE)--April 16, 2003

Building on the company's history of performance leadership with the Itanium(R) processor family, HP (NYSE:HPQ) today announced commercial benchmark results on the HP Server rx5670 using the next generation of the Intel(R) Itanium 2 processor (code-named "Madison").

Established on the SPECweb99_SSL(1) benchmark and the two-tier SAP(R) Standard Application Sales and Distribution (SD) Benchmark(2) with servers running HP-UX 11i, the new leading results reinforce the outstanding commercial performance of HP's systems based on Itanium 2 processor 6M (Madison) running at 1.5 GHz with 6M L3 cache, due to ship this summer.

Running key industry applications from SAP AG, Oracle, Zeus and others, the HP servers outperformed IBM Power 4+ based and Sun UltraSparc III-based system marks on the tests by 40 percent to 400 percent(1-6).

"These initial tests of enterprise resource planning and secure e-commerce transaction workloads are solid proof that users with the most advanced and data intensive workloads will see superior benefits from HP systems and the next-generation Itanium processors," said Mark Hudson, vice president marketing, HP Business Critical Systems.

The latest results also top by more than 40 percent the previous four-processor highs on both benchmarks, held by currently shipping Itanium 2-based HP servers.

SPECweb99_SSL Benchmarks on HP-UX 11i

HP posted world records for both four- and two-processor systems on the SPECweb99_SSL benchmark(1), which measures performance for companies executing high volumes of secure transactions over the Internet.

Working with Zeus, a leading provider of Web server infrastructure(3), the HP Server rx5670 achieved a four-processor result of 3,344 simultaneous connections running HP-UX 11i(1). This surpasses IBM's four-processor Power4+ based result by more than 65 percent and Sun's four-processor UltraSparc III-based result by more than five times(1).

The HP rx5670 also posted the fastest two-processor result with 1,750 conforming simultaneous connections(1). This surpasses Sun's 1,200-MHz UltraSparc III-based result, which uses Sun Crypto Accelerator 1000, by more than 70 percent(1).

These world records were achieved on the HP server using the Itanium 2 processor 6M (Madison) at 1.5 GHz with 6 MB of L3 cache.

SAP Standard Application SD Benchmarks on HP-UX 11i

The HP Server rx5670 achieved the leading four-processor result on the two-tier SAP Standard Application Sales and Distribution (SD) Benchmark(4) of 860 SAP SD Benchmark users(2) with 1.97 seconds average dialog response time. With 215 SAP SD Benchmark users per processor, the server also demonstrated leadership on a per-processor basis, a result approximately 40 percent greater than IBM's eight-processor Power4 result(5) and nearly three times the per-processor result posted by a Sun eight-processor SPARC III system(6).

The HP rx5670 based on the Itanium 2 processor 6M (Madison) achieved these outstanding results running HP-UX 11i, SAP R/3(R) Release 4.6C, Oracle(R) 9i Database Release 2 and HP StorageWorks Virtual Array 7100 with fifteen 18-GB disks.

The SAP SD Benchmark measures six different e-commerce transactions, including order creation, changes to existing orders and invoicing.

HP servers and workstations running the Intel Itanium 2 processor 6M are planned to begin volume shipments in mid-2003. In addition to HP-UX 11i, the HP servers based on Itanium 2 processors also support Windows(R) and Linux, providing HP's customers with an infrastructure that can flexibly adapt to their multi-OS workload environments.

"These superior benchmark results are further proof that Itanium 2-based HP systems, with their leading performance, price/performance and multi-OS advantages, continue to set a standard that no other 64-bit architecture can match," said Hudson. "The fact that these results were achieved with pre-production 'Madison' Itanium 2 processors is proof of the superb quality of the Intel design and another milestone on the way to delivering more industry-leading Itanium 2-based systems from HP this summer."

More information on the HP Server rx5670 and its industry-leading benchmarks is available at http://www.hp.com/go/rx5670.

About HP

HP is a leading global provider of products, technologies, solutions and services to consumers and businesses. The company's offerings span IT infrastructure, personal computing and access devices, global services and imaging and printing. HP completed its acquisition of Compaq Computer Corporation on May 3, 2002. More information about HP is available at http://www.hp.com.

(1) Competitive benchmark results reflect results published on http://www.spec.org as of April 15, 2003. The latest SPECweb99_SSL benchmark results are available at http://www.spec.org/osg/web99ssl/.

SPECweb99_SSL: HP Server rx5670: 3,344 conforming simultaneous connections using four Intel Itanium 2 1.5-GHz processors.

IBM eServer P630: 1,988 conforming simultaneous connections using four Power4+ 1.45-GHz processors.

SunFire V480: 568 conforming simultaneous connections using four UltraSPARC III 900-MHz processors.

HP Server rx5670: 1,750 conforming simultaneous connections using two Intel Itanium 2 1.5-GHz processors.

SunFire 280R: 1,008 confirming simultaneous connections using two UltraSPARC III Cu 1,200-MHz processors.

(2) Running the two-tier SAP Standard Application SD Benchmark, HP achieved this result based on the following configuration: four 1.5-GHz Itanium "Madison" processors with 6 MB of L3 cache and 24 GB of main memory. The measured throughput was 86,330 fully business processed order line items per hour and 259,000 dialog steps per hour with an average CPU utilization of 99 percent for the central server.

(3) More information about HP's Zeus-based Secure Web Server is available at http://www.hp.zeus.com.

(4) The SAP benchmarks were certified by SAP AG and fully comply with the guidelines issued by the SAP Benchmark Council. Results referenced are current as of April 10, 2003. More information on SAP Standard Application Benchmarks is available at http://www.sap.com/benchmark.

(5) 1,220 SAP SD Benchmark users for IBM's p650 server running R/3 Release 4.6C, DB2 UDB 8.1 and AIX 5.2. The configuration used eight Power4 processors at 1.45 GHz with 8 GB central server memory. Average dialog response time was 1.95 seconds at 122,670 line items per hour and 368,000 dialog steps per hour.

(6) 600 SAP SD Benchmark users for Sun's SunFire v880 server running SAP R/3 4.6C, DB2 UDB 7.2 and Solaris 8. The configuration used eight UltraSparc III CPUs at 900 MHz with 32 GB memory. Average dialog response time was 1.96 seconds at 60,330 line items per hour and 181,000 dialog steps per hour.

Note to Editors: Intel and Itanium are trademarks or registered trademarks of Intel Corporation in the United States and other countries. SPEC and the benchmark name SPECweb99_SSL are registered trademarks of the Standard Performance Evaluation Corporation. Oracle is a registered U.S. trademark of Oracle Corp., Redwood City, Calif. SAP, R/3 and all other SAP product and service names mentioned herein are trademarks or registered trademarks of SAP AG in Germany and several other countries. Windows is a U.S. registered trademark of Microsoft Corp.

This news release contains forward-looking statements that involve risks, uncertainties and assumptions. All statements other than statements of historical fact are statements that could be deemed forward-looking statements. Risks, uncertainties and assumptions include the possibility that the market for the sale of certain products and services may not develop as expected; that development and performance of these products and services may not proceed as planned; and other risks that are described from time to time in HP's Securities and Exchange Commission reports, including but not limited to HP's quarterly report on Form 10-Q for the quarter ended January 31, 2003, and subsequently filed reports. If any of these risks or uncertainties materializes or any of these assumptions proves incorrect, HP's results could differ materially from HP's expectations in these statements. HP assumes no obligation to update these forward-looking statements.